Hacker News new | past | comments | ask | show | jobs | submit login
Ask HN: Idea Sunday
305 points by rokhayakebe on Apr 13, 2014 | hide | past | web | favorite | 441 comments
A HN experiment. Every Sunday, a thread will be started to share product ideas. Why? Because many people have ideas they will simply not have the time to implement, and many need product ideas to work on.

If you think this thread should be started only every other Sunday or monthly, please state it in your comment.




I'd like to see a site that focuses on the construction of logical arguments supporting various political and non-political views.

I'd like for the system to allow for continued refinement of arguments on various topics and to allow for various branching of thought. That is, I'm not looking for a democracy that rewards majoritarian thought and punishes less popular but logically consistent argumentation.

As an end product, if I'm discussing, say, gun control with someone, I should be able to give them a URL: http://www.xyz3323499.com/crusso/gun_control/ and all of the arguments and important information that I have upvoted, added to, or edited would show up.

Most of this stems out of my frustration with the tedious nature of arguments on the internet. New stories stir up old arguments. Forums like HN tend to rehash all those old arguments with often little thought for the mountains of arguments/evidence already out there. Rather than have new stories refine or even change arguments and their conclusions, most people on the internet just chase their tails in circles.

I'd view this kind of site as a "let's stand on each others' shoulders and provide our shoulders to others" kind of work. With a decent web of trust, I should be able to look at the arguments of the most "respected" members of the site and mix-in their arguments with my own.

I think that a sufficiently complex web of trust would allow for fundamentally different viewpoints on a given issue to thrive because the spheres of supporters of those viewpoints have different value systems. Even with those different value systems, it should be possible at some point to pick out the thought leaders across different spheres that transcend typical Republican, Democrat, Liberal, Conservative, Theist, Atheist labels with arguments that are acknowledged across the board as being internally consistent and factually based.

Yeah, I realize that I've rambled -- but if someone finds value in the idea and runs with it, I'd love to see it out there.


I've explored similar ideas in the past. One possibility I explored was breaking down claims like 'Guantanamo Bay should close' into moral claims plus factual claims, eg 'Guantanamo Bay practices torture' plus 'torture is wrong'. This could be further broken down until you had a set of moral axioms and a set of basic pieces of factual evidence, and a set of syllogisms linking them together.

Different people would indeed have different value systems and also give different degrees of credence to different sources. So the goal would be to end up with a crowdsourced tree of moral & factual claims, and individual users could then decide which claims they agreed or disagreed with to come to their own conclusion on the topic.

(Look up 'argument mapping' for academic exploration of the concept).

My thinking now is that most online political arguments online are really about tribalism, not finding the truth. If you did have users who cared about the truth (and who could consider other viewpoints), you wouldn't need a fancy product: a simple web forum would suffice.

But do drop me an email at isaac@i.saac.me if you'd like to chat more.


I agree. Breaking down a debate into arguments which have moral and factual premises will be useful. The logical validity of the argument can be automatically enforced. The system wouldn't allow you to make an invalid argument.

Showing people exactly where they disagree would be a huge advancement. I think there are plenty of people who care about the truth and also care to have a consistent worldview in terms of morality.


For me and you, this is true. Unfortunately, INSERT_MAJOR_NEWS_NETWORK is proof that the vast majority of people don't give a shit about logical validity.


I think that the success of sites like Snopes indicates that a significant percentage of people DO care about facts.

We'll never have a situation where 100% of people are 100% fact driven in their belief systems, but maybe the needle might move in the good direction if more tools were available to help the transition.


Maybe something like http://www.riyarchy.com/issue?title=Euthanasia or http://www.crystaltower.co/welcome ?


Fascinating, thanks for the pointers.

I'll have to look at them further, but what I think they might lack from what I posted is a notion that different users will value some gray area arguments differently. The system should preserve those differing value systems in the arguments constructed and use them to build upon the webs of trust. For example, user bob may find it axiomatic that "healthcare is a right". User jeff will typically downvote arguments that assume that right and not want to build his framework on top of it.


I love this idea.

I don't think most people -myself included- would be able to accept the rational arguments. (There is probably research showing that people's beliefs become more entrenched if you show them evidence to debunk or otherwise convert them).

Using gun control as an example. I have a position. I'm in a discussion with soomeone. I think I'll use evidence to persuade them, so I visit your site. I find that the argument pretty much demolishes my point. I thus ignore your site on that occasion, complaining about bias and lack of use of the site hy right-headed people, and I add my tedious arguments to my discussion.

Or I find that the argument supports my position, in whichcase your sote is brilliant and the Internet is a marvel so I post a link, which the person I am talking to ignores.

But it is still a great idea and you're right that some subjects (abortion; gay marriage; drug control; gun control; Israel / Palestine; circumcision; etc) get discussed a lot, with repetitive and circular and eventually boring arguments.

Seeing how well, or poorly, Wikipedia has handled hot-button topics points out some difficulties of finding "accuracy" when people are not rationally calm headed.

It would be great to see it though.


I love this idea.

Thanks, it's one of those ideas that has been in the back of my mind for a while, but it's not an easy idea for me to even convey properly. I thought I'd just make a best-effort at it and see what HN thought of it. The feedback has been really interesting and I have a bunch of URLs for somewhat similar concepts that I plan to spend some time on.

I don't think most people -myself included- would be able to accept the rational arguments.

I can't speak for everyone, but I can name several major political/religious positions I've had in my life that I've changed my mind on because of discussion/debate and research.

   *I find that the argument pretty much demolishes my point*
   *Or I find that the argument supports my position*
Actually, I'd think that you could find arguments in support of either. I don't envision this kind of site as being the terminator of all debate, since different perspectives would be allowed to thrive in different spheres based upon the complex web of trust concept.

What it would allow you to do is say to someone on some random forum, "Look, I'm not interested in taking a trip on the merry-go-round with gun control in a Youtube comment. Here's the URL to my honest assessment of the issue, if you're wanting to do something more than just grandstand and create a bunch of strawmen about my arguments, there's a starting point for you. If you have a reference for your position, I'm more than happy to look at it. What's your URL?"

You can't reach everyone with logic and data, but I'd love to see a tool set that encourages people to explore their own positions in an environment that promotes really thinking about those positions rather than just going in with shields up and phasers set to kill.


This is along the lines of structured arguments http://en.wikipedia.org/wiki/Argument_map

Doug Engelbart the creator of NLS and famous for Mother of All Demos talks about such systems http://youtu.be/xQx-tuW9A4Q and http://youtu.be/VeSgaJt27PM etc.

The general principle is we won't make much progress discussing complex issues by simply recreating the print world with increased convenience of a computer like we do now. The discussions have to take a fundamentally different form, possible only on computers, such as semantic argument maps, etc.

Of course his perspective comes from 1950's when educated people had civil discussions that solved problems, or at least that's the modern nostalgia. He imagined a new class of "knowledge workers" moderating information. Instead we have voting rings and content farms.

If anything, a structured argument map could shut down much of the attention seeking drivel surrounding every political/complex system issue.


hear hear. Won't happen in my lifetime, but if something like mathematica or ipython workbooks could become a common way to present an argument in a public debate then much would have been won.

btw: For presenting complex issues, journalism is dead. For uncovering hidden truth, not necessarily very complex ones, it's still very much alive.


(inserts tongue firmly in cheek) I think there would be more money for a site that helps you argue your position from available data, the truth be damned.

I'm thinking "How to Lie With Statistics"[1] as a service - tell it you want to, say, show that guns are becoming safer, or that higher taxation causes obesity, and it will produce a nicely manipulated graph containing only real figures from readily available datasets, ready for feeding to a journalist.

[1] http://en.wikipedia.org/wiki/How_to_Lie_with_Statistics


> I think there would be more money for a site that helps you argue your position

Let me toss in some anecdotal evidence in support of your suspicion. The only "debate" site to hold my attention for more than a first visit is one that focuses on curating data to support arguments about certain issues (the curation is manual AFAIK). I initially stumbled across it when debating health care reform:

http://healthcarereform.procon.org/view.resource.php?resourc...

Pay particular attention to to "Projects" sidebar on the left. It's a goldmine of relevant quantitative information and the arguments surrounding their interpretation. I actually don't think it's cynical to expect that this is the right way to approach the "debate market." Discovering data in relevant "digested" forms is a huge friction point. It should be central to the political debate process, it isn't, and anyone who can reduce the friction will be doing a huge service to humanity.


This is an interesting idea. I am consulting in data analytics and I get two type of clients. One where client hasn't made decision and looking for data analysis to guide them. Second where client has already made up mind and looking for data to convince others of their decision. How to lie with statistics can easily fulfill need of second type of client.


What I've learned is that complex political issues -- essentially anything that isn't a universal "no-brainer" -- comes down to tradeoffs.

In fact, I've found a good political test to see if the person I'm speaking to can articulate a valid criticism of their own position. I.e., if they're for universal healthcare can they state why one would oppose it, without calling into question the opposition's character.

If they can, you can have a fairly thoughtful discussion/debate on the topic.

If they can't, then there's no point discussing things.


I use a similar methodology when teammates disagree on a design point, have them argue the other side. In nearly every case the dispute is settled.


I recently launched a webapp, CapitolBells.com, that is sort of a HN for arguments supporting or opposing legislation in Congress ("crowd-lobby Congress"). A friend of mine runs a project called Madison, which in essence is GitHub for legislation, allowing for various branching debates and amendments to bills. We've often thought it would be cool to combine the two ideas.


Edit for down-voters: I'm not trolling/insulting (people who post on) internet boards. I'm hypothesizing that on most topics (esp. politically charged topics, which tend to rely on hard-to-assess factual claims), a high-quality argument for or against an opinion might require years of experience, and that often communicating that experience often requires journal-article-length explanations without drawn-out review processes governed by other experts. This is not an unreasonable possibility.

And if this is true, then the parent's idea might not be feasible. If it isn't true, then the parent might have a feasible idea. So... perhaps you could comment explaining why you're down-voting? I honestly don't understand.

(Also, the first sentence is just true... the venues I mentioned are nearly universally more informative than a comment board... srsly, not trolling.)

Original post:

If you want high-quality commentary all you need to do is get off the internet comment boards. These outlets exist. They are generally called journal articles, policy briefs, court opinions, etc.

> Rather than have new stories refine or even change arguments and their conclusions, most people on the internet just chase their tails in circles.

I think there's something deeper going on than a short memory span.

At some point (usually pretty quickly), you have to start either 1) making moral judgments, which is a whole branch of philosophy or 2) analyzing non-trivial factual claims.

To do this, you need people who are trained in science or well-read in philosophy/political science, depending on the topic. At that point, might as well just read journal articles in relevant fields.

The quality of argumentation on the internet might get a little better than it is today with good tools, but I don't think it can improve significantly. At some point you're asking for actual quality research (i.e. work), as opposed to armchair sophistry (i.e. fun).

> Even with those different value systems, it should be possible at some point to pick out the thought leaders across different spheres that transcend typical Republican, Democrat, Liberal, Conservative, Theist, Atheist labels with arguments that are acknowledged across the board as being internally consistent and factually based.

Maybe on some technical topics. But even academics and policy analysts often disagree along ideological lines.

> as being internally consistent and factually based.

These are both fairly low bars.


I upvoted you to counter downvotes. I think the downvotes are for the offensive part of "get off the internet." How dare you suggest such a heresy?

I agree that complex system discussions must be driven by people educated in the ways of complex systems, possibly from perspective of economists, justice system, etc.

Of course digging through the hieroglyphics around the issues becomes difficult for laymen, even those educated in the relevant topics. And it does become more work than fun. But many online discussions are at the level of work already. People start them for "fun" then get frustrated for being wrong, downvoted, misunderstood etc. There is an addicting effect and it turns from "fun" to work of salvaging reputation, revenge, etc.

Also content farms release misleading, throwaway information that gets attention. For them it is work and often very not fun.

Their efforts could be channeled to better ends by new software.


Right: being able to read and write an argument doesn't make people acceptably qualified experts on a subject.

Compare this to Yahoo Answers: Somewhere where everybody has an opinion, but nobody is expert enough to give a straightforwardly correct answer to the most basic questions.

If your target is to have more intellectual questions, I don't think it follows that it will be straightforward to attract more informed opinion.


I like the train of thought.

"Voting" on the internet definitely added a lot but it created a bunch of hurdles: majoritarian thought, bubbles and averaged opinions.

Review sites like rotten tomatoes are a good example of the last one. The average opinion about a movie is nowhere near as good as asking a single person who shares my tastes in movies. A bubble would be much more useful here. I'm not sure where I'm going with this.

I think letting ideas branch might let controversial discussions get objective. Complicated can be mistaken for subjective. It also sounds like the kind of fun that can draw people in.

I really like the idea.


I'm kind of waiting on Facebook to do this. Create a voting engine based on people similar to me for services like Rotten Tomatoes and Netflix. The "other people similar to you liked.." should be far more ubiquitous.


I appreciate the idea, but it needs some reworking.

In particular, I see two big flaws.

(1) The essence of the use of this site would be that one chooses a position, and then finds arguments to support it. In other words, "I've made up my mind; now show me why I'm right."

Now suppose I'm the person discussing an issue with you. Why should I pay any attention to some page chosen based largely on the assumption that I'm wrong?

(2) You say "... I'm not looking for a democracy that rewards majoritarian thought and punishes less popular ...." And then you apparently want to use something along the lines of up/down voting to judge arguments. That's not the way to get what you say you want.

In short, the idea, as stated, strikes me mostly as a tool for increasing polarization.

More beneficial would be a site that simply provides information to help make decisions on whatever issues are of interest. To go with the gun-control example, you could have links to -- or summaries of -- studies on the effect of gun control. You could have comparative data on rates of various crimes in locations with different gun laws. Etc.

How do deal with (2) is a harder problem. The plague of voting-based-on-agreement that makes sites like HN & Reddit so much worse than they could be, is something I wish more people were trying to address.


1) I'm sure some will approach it that way. Some of those will continue with it that way. Other people will learn from the ongoing debate and incorporate new learning in their positions. They will learn logical argumentation and some facts because thought leaders will emerge who leverage logic and facts. Those thought leaders can be incorporated into webs of trust as preferenced by individual users.

2) I'm not looking to decrease polarization, necessarily. If anything, different echo chambers would necessarily form as like-minded sets of users would agree with and upvote similar arguments. If user sammy upvotes weak creationist arguments made by weak creationist supporters, fine. Within the region of the web of trust of weak creationist arguers, sammy may be king. The site should allow him his due, but when sammy goes to upvote a new argument that crusso is following, it should affect the ranking/view of that argument little if at all -- since crusso tends to not upvote weak creationist arguments and creationist argument supporters.

Although mechanics that allow for increasing polarization sounds bad (I hear it), the conventional approach of decreasing polarization is to beat the crap out of outliers with monolithic, single-dimension voting systems that reward populism and emotional rhetoric.

I'd rather let some polarization thrive because different people have different value systems... but what I would envision as being a "good" for a site like this is to allow people to both have their polarization AND to allow them to transcend their echo chambers by viewing the arguments of other thought leaders in other echo chambers using similar tools and metrics for evaluating arguments.


Perhaps an implementation of the idea could reduce arguments of a political thought into an axiom, e.g. (this is probably wrong) abortion should be legal....<insert a tree of arguments here>..., assuming individual liberty is the most important quality of a society, vs abortion should be illegal....<insert a tree of arguments here>..., assuming maximising human life being the most important quality of society.

In that case you could see where you fundamentally disagree with a political idea, and accept the difference quickly without rehashing past arguments.


I have some ideas for improving the efficacy of online voting systems, but starting a new online community is hard. Even just starting a subreddits is hard, and I imagine that to be lower-friction than creating a stand-alone site, given Reddit's large existing user base.

If you have an online-community that you run and would like some ideas to combat vote-based-on-agreement, here are a few that I've thought up or come across:

(Not claiming that these strategies are ungameable).

* Limit the number of upvotes a user can give per day, thus making votes more valuable.

* Force users to space-out voting, allowing only 1 upvote every 5 minutes.

* Vote value: Votes are valued more the less you do it over a certain period - so if you only upvote once per-day, that post gets more points than if you upvoted everything in /new. Choice: Do it hidden on the backend, or could be interesting; do it in a user-configurable format. Give users, say, 100 points per-day and are allowed to upvote transparently - you can give one thing 100 points or four things 25 points.

* Time-delayed voting: votes are only allowed after a certain amount of time has passed since you clicked on a link. A well thought-out and reasoned article should take at least a couple of minutes to read, so I can upvote that after I read; I'm not likely to wait around to upvote an image-macro with fewer than a dozen words unless it's really special.

* Karma cap, both negative and positive: Limit max positive karma to something low, say 50. That's high enough to see that a user isn't brand new but isn't trivially attainable. That way, it's fine to get to the cap, but posting inane one-liners attempting to reap thousands of points for a single comment isn't so tempting. Limit negative karma to something lower, like -100 to dissuade trolls, while still showing that a user is disagreed with. Hopefully that encourages discussion for the sake of discussion.

* Meta-voting: You vote on how many points a post got, so if a pithy but well-timed comment that says "lol butts" gets a lot of points, you can vote to say that the post didn't deserve those points. Make this a limited power granted to a select (but changing) group.

* Make downvoting cost karma: If it cost karma to downvote, that could be a higher bar to drive-by-disagreeing.

* Not sure if it's a good idea, but turn the script around. Instead of trying to avoiding the Pepsi-Blue Advertising syndrome, charge money for votes transparently adding points to a post or comment. As users can still downvote items, it becomes expensive to elevate a bad piece of advertising. Maybe extend this to users by allowing users to receive money for lots of highly-rated comments. Or perhaps use Dogecoin as your currency of choice as the real-world value of 1 Doge is so low that it's effectively free in small quantities, but not actually free.

I don't think allowing the masses to vote on things is a bad idea, but it's an idea that needs some refinements for best effect.


I had an idea for something like this during the 2008 Presidential Elections. There was so much hearsay and misdirection that I did not know where to turn when trying to decide who to vote for. My idea was more of a forum where you could somehow remove the emotional human element out of the discussions and a moderation system that would not allow posts that could not be supported by credible sources. I remember I came up with a cool name, checked Godaddy.com, saw it was taken, and abandoned the idea. I think having a repository of credible facts to turn to when you need to support a philosophical argument would be awesome. Kind of like a Snopes.com meets Wikipedia.


I'm working on a project called Whybase that does exactly that: http://whybase.com


The talk.origins archive (http://www.talkorigins.org) attempts to do this for all of the arguments posed by creationists against modern evolutionary theory.

Unfortunately, there are still creationists. Even when there is overwhelming scientific evidence in support of one position, people will still hold another.


Check out http://idebate.org/debatabase I knew it as debatepedia a while ago. Choose a topic and click on a subtopic. The old version had a argument map style debates. There is already a mention of argument maps in one of the comments which is another great tool.


How would you respond to someone with an opposing point of view on an issue you're passionate about replying to one of your comments with a link like that?

It's an interesting idea, but I think the conceit is that other people would actually read your arguments at that link. In my experience, that's not true.


I think every attempt at an argument should be followed by an attempt at a counter-argument.

For example if I cite a study as a reference, a counter argument could be the methodology used to gather the data

Edit: @harpastum made an excellent suggestion of existing sites. Disregard this


http://lesswrong.com/


Less Wrong explicitly doesn't talk about politics (or at least tries not to), so it's not the right venue for the sort of thing OP wants.


Its been around for years

http://argumentrix.com/wiki/Main_Page

EDIT: opps looks like its dead.

shouldn't be to hard to clone


Illustrated: http://geniscarreras.com/philographics/


You might be interested in this http://www.procon.org/


A simple take on your idea... http://wrangl.com/


Argh, my idea! I really need to start working on this, it's been bouncing around in my head for too long


Ha, same here. Too many other things that take priority at the moment though.


In order for such a project to make sense, one must ask himself if he believes that there is one and universal truth that stands above human and physical laws. If there is such a thing, then there is a point in pursuing the truth, because truth gives the correct conclusions. If not, then there's no meaning in pursuing something that is relative.


Of course there is one truth. 1+1=2, Global warming either is or isn't real, gun control either does or doesn't reduce homicide rates, etc.

Aumann's Agreement Theorem suggests that two rational people can not disagree on a subject if they both fully understand all the evidence and arguments. All disagreements stem from some combination of misunderstanding, ignorance, or irrationality.


That's only true for factual statements. It's perfectly possible to reach different conclusion starting from different moral premises.


Then there is still only one truth and they can still come to an agreement on statements like "gun control is good if your goal is to reduce homicide rates" or "gun control is bad if your goal is to maximize personal freedom", etc.

Then it's not a matter of disagreement or coming to different conclusions, it's just having different values.


I think this works better. As someone said above, it comes down to trade offs. Arguing a point with a specific goal to work towards leads to better discussions than trying to answer questions like: "is x good or bad?"


Some questions of predicting a very distant future can't have an accurate answer today. But if we can simulate our world with a computer, try different things and see where they lead in the near future, we can answer most of our complex system/politic/economic questions.

But such a computer would obsolete humanity.


Calendars & Events as a Service (aka "Stripe for Calendars")

I'm sick of dealing with scheduled events, timezones, future background jobs, etc. in almost every app I build. I want a service with an elegant API and nice libraries that would let me:

* Create multiple calendars (it would have an ID I could then save with my user model or whatever)

* Add events to a calendar by just passing it a timestamp and other metadata

* Have webhooks that get called any time an event starts/stops (possibly also X minutes before)

* Nice API calls for getting all a calendar's events, checking free/busy times, specifying time zone output and strftime formatting, etc.

* A JS library for outputting pretty calendars to an HTML page

* An additional cron-like feature for each calendar allowing you to define recurring webhooks

If anyone wants to give something like this a shot, let me know. I'll be your first customer, partner or angel investor.


This is interesting. I had an idea a few years back that was a variation of this as a consumer/small business service, rather than a developer one.

The idea (roughly) was predicated on the belief that much information that needs to be communicated in the general sphere of "local" is in the form of calendar events. Everything from community meetings to wine tastings to school events, soccer games, carpool times, and so on, can be communicated in the form of calendar events and their metadata. So the concept was to do a clearinghouse for discovering and subscribing to events and event feeds (which would auto-update when changed and so on) that also made it easy for publishers of events to enter them in a single format (calendar feed, email-to-event, etc) and have them translated into whatever formats were preferred by the end-users.

EDIT: wondering if i just described Upcoming.org...


I find this interesting. Calendars are fun, until you have to program them :P


Just out of curiosity, isn't this already possible with the Google Calendar API?


I haven't looked that deep into the most recent version of G's calendar API. I don't think they have webhooks though.


Actually they do in v3 https://developers.google.com/google-apps/calendar/v3/push


Does it allow you to create Calendar as a data object whenever you want to create one or is it just one calendar associated with one google account ? Because what if I don't want to create google account for my users , just get the calendar functionality.


You can create multiple calendars but they are always associated with a user's google account. You don't want Stripe for calendars, but rather Balanced for calendars. A whitelabel calendar API where you user's dont have to sign up for another service in order to use yours. Correct?


Correct!


I'm working on this (a subset of these features at least) right now. It was going to be an internal tool for my company, but having seen this interest makes me think I might try and go the SaaS route...


This is definitely something I would use. I hate messing with date/time. Timezones, zillions of different date formats, etc. all make dates and times a pain. A slick "Stripe-like" way to handle all of that would be awesome.


I'm considering building something like this. Just to be clear, is converting user input into timestamps part of the problem (and if so, do libraries like http://momentjs.com address it)? Or are you strictly referring to the scheduling aspect of it?

Is this something you would pay for? How much?


Interesting. I was thinking about a very similar calendar API last night. Mostly from a scheduling standpoint. But the webhooks part is a cool addition. Is something similar now possible with zapier?


Doesn't Zapier have to act on some outside event? Or are you saying combine Zapier w/ Google Calendar to solve the webhooks problem?


Sure, you could prob hook it up to gcal. But I thought they may have a timer event as well.


This is so true. Dealing with time-related development features is one of the most annoying things I deal with.


There is this one: http://php.brickhost.com


i'm interested. any other devs want to collab? anon1@alumni.stanford.edu


We like experiments and this is a good one that goes to HN's core. Also, "Idea Sunday" is a great name.

I don't see why there shouldn't be a thread like this every Sunday, as long as the quality is good. If there aren't enough new good ideas, we can reduce the frequency. Consider yourselves challenged!

If it is to become a regular thing, we'll ask the whoishiring account to post them. That's the only account allowed to make recurring posts like this, which solves the problem of users competing to do it for the karma. Maybe rokhayakebe can keep posting these for another few weeks and, if it's still going strong, we can semi-formalize it then.


What I'd love to see is some kind of "meta" HN where all the ask HN, show HN and similar threads are posted. For two reasons: 1) those "meta" threads might gather to a different audience and 2) many people are more interested in normal threads, and many are interested in the meta.

One could just go ahead and create a site for that, but obviously it wouldn't have the same traction as HN, and 99% of the value of HN is in the people posting here.

I do realize that it will not get implemented, but maybe consider this :) Maybe instead of creating a new section of the site, make those meta thread listen under an url like https://news.ycombinator.com/meta, so that one interested could see the latest threads without much effort.


And one for bitcoin! And one for nodejs! And x, and y and z! You know what, let's put them in a subdirectory, like so: /h/bitcoin or /h/meta. And we call them sub-HNs!

...


Although a sub-HN approach sounds appealing at face-value, it's likely to fragment the site content a little too much.

One of the things I appreciate most about HN is its ability to act as a filter with single point of convergence. I'm not sure I'd want to sacrifice that (personally), probably nothing more than for 'fear of missing out'.


Pretty sure he was being sarcastic, referencing reddit, subreddits, and the /r/ url schema.


There is an experimental Firefox add-on that attempts to sort out the ask HN, show HN, who's hiring and polls in their own tabs. https://addons.mozilla.org/en-US/firefox/addon/hacker-news-r...


I think it would be interesting to have an app that determines the optimal times to stop for gas during a trip. Especially when you are traveling across state lines, gas prices can vary a pretty decent amount.

The simple version:

- Once you're under a certain amount of fuel (1/3 of a tank?), you open the app, and tell it you need to get gas within the next n miles. The app checks gas prices along your route (does gasbuddy have an api?), and tells you which gas station you should stop at.

The complex version:

- The app connects to your car with an OBDII device via bluetooth (they are pretty cheap), automatically watches how much gas you have left, and determines the optimal place to refuel once you are under 1/3 of a tank. Basically, you would be able to open the app at the start of your trip, and it would alert you a few miles before the best exit to get gas.


Commitment device app that works on a dead man's switch. The idea is to use the discipline you have now to discipline your future self.

eg1 - Enter your weight every month. If you go over a predetermined weight, it does something punitive. Posts embarrassing pictures online, emails your mom, donates $50 to the Klu klux Klan.

eg2 Create a todo list with due dates, goals or somesuch. Failing to complete them does something punitive.

eg3 Create out a challenge (exercise 5 times per week). Failure to complete challenge does something punitive.

The key is to get the psychological component right. I think if someone's daily task is 100 or more pushups they are more likely to fraudulently push a complete button than they are to enter a fraudulent number of pushups.

Could be a fun project.


I use http://beeminder.com/ for this. For my weight, I have a Fitbit Aria scale -- I weigh in daily, it uploads my weight to Fitbit and the data is pulled in by Beeminder. So there's no friction in getting data into the service (I just blearily step onto the scale every day) and there's no way for me to lie, even if the thought of financial loss tempted me to.


You might enjoy this story / podcast: http://freakonomics.com/2012/02/02/save-me-from-myself-a-new...

One guy decided that he wanted to get healthy. So, he made a list of prohibited foods and actions and wrote a check for $750 payable to Oprah Winfrey (he dislikes Oprah for some reason).

He told his wife, "If I do anything on this list in the next 30 days, send this check to Oprah, no exceptions."


I think I may have gotten the KKK idea from freakonomics podcast, might have been this episode. The Oprah things sounds familiar too.


Our rebuttal to the anti-charity idea: http://blog.beeminder.com/anticharity (In short: it's socially inefficient!)


Beeminder, as others have mentioned, handles #1 and #3 pretty well (minus the donation to the KKK part - the money for failing goes to Beeminder).

I built GTBee[1] using the Beeminder API which is pretty much #2. Again, the punitive thing is taking your money, starting at $5 and escalating from there.

This isn't to say that it wouldn't be a fun project to build things that do this though!

[1] https://itunes.apple.com/tc/app/gtbee/id779525180?mt=8


Huge thank you to everyone pointing out Beeminder (I'm a cofounder).

Here are all the other such apps we know of: http://blog.beeminder.com/competitors


I'm impressed you put that post up & link to it here.

:)


Isn't this what beeminder.com does?


Yes, it is. (The specific punishment it uses is charging you money – more money each time you fail). Clickable link: https://www.beeminder.com/


See also, Gym Pact [0]. An app that lets you earn rewards (versus payout cash) for hitting gym goals. Similar idea of adverse motivation.

[0] http://www.gym-pact.com/


wasn't this stickk.com?


Lower ligature risk USB charging cables. These would be used by mental health hospitals to allow (risk assessed) patient to charge their devices without introducing ligature risk.

To go along with this:

Bulk USB charger. Imagine a 20 port USB hub, but this has no USB connectivity. It only has USB power. It would be kept in a secure place in a mental health ward and would allow them to charge patient's devices. This would mean that patients get electronic devices but without having access to cords (ligature risk) or mains plugs (self harm risk).

(Take a UK 3 pin plug. Place it on floor with prongs up. Jump off bed onto plug with weight on one foot. That's an unpleasant injury. Removing that risk is useful).

Advantages:

You don't need to PAT test each charger coming into the hospital. You reduce (slightly) fire risk from bad quality chargers.

You reduce ligature and self harm risk.

You allow patients access to electronic devices which has some "safeguarding of liberties" benefits.

Disadvantages: selling to the NHS is possibly hell, and selling electrical equipment to the NHS is possibly even harder.

You're creating a vigh value stash of easy to steal equipment. MH hospitals already have lots of IT and stealable medication so they should be able to keep it safe but maybe a ventilated safe would be part of the package.


This is definitely not a short-term solution, and it might not be practical even in mid-term, but wireless charging sounds like a perfect solution for this kind of environments.

The hospital rooms could have special "charging" spots on the floor/build into the furniture or just a hole in the wall. The patients would charge their devices simply by placing them on the charging spot/shelf. Now there are no potentially dangerous cables or plugs, and all charging can be done without devices leaving the patients' reach, reducing the risk of theft at the same time.

There's already an international standard for wireless charging. The only problem is that it works only with a couple of the current phones, and it's not very likely to appear in low-cost low-end devices that the patients of the mental hospitals most likely to use. But if it ever gets a wider adoption, this would solve a lot of problems...


Would something like this fix that issue? https://www.indiegogo.com/projects/chargekey-for-iphone-and-...

I backed this on Indiegogo and received it a few weeks ago. It is quite small.


That's what I'm thinking. Short cable, and rubberize the metal ends so they can't be used for cutting.


Wouldn't it will be simpler to have replaceable faceplate with different connectors (a cradle) in each room? Basically, idea is similar to having power buttons/plug/ethernet connector on the wall. Instead you have horizontally mounted faceplate in a cabinet, bed headrest, or any horizontal surface with appropriate connector where device can sit and be charged. There are not too many USB charging devices that will be used in a mental hospital by patients. And depending on the type of device swap out the faceplate with appropriate one.

The best way to avoid ligature and self harm risk is to not have any cable or sharp objects around. Also, this solution will avoid the privacy risk when patient device is taken away for charging by staff members.


The anti-ligature cord I have seen works on the basis that as soon as too much force is applied, it breaks into pieces:

http://www.hoyles.com/acatalog/anti-ligature-sanipull-pull-c...

(scroll down for the examples)

Perhaps it would be possible to incorporate the USB charging wires inside that and still have it work?


Re: reducing ligature risk. Would having a rigid cable (i.e. hard/unbendable metal shell) that's short be a reasonable solution? Similar to the indiegogo project posted in a sister comment, but could be slightly longer and cheaper to manufacture.


wouldn't it just be a matter of using a shorter cable like 4" or less?


Well, the cables have to be really robust so nothing can be extracted from them that could be used for cutting.

Some people can cause surprising damage with a bobby pin so having robust, potted, cable shells would be important.


Or a strong magnetic connector every 10 or so inches to allow you to build a cable as long as you like? (I have no idea if that's actually possible)


Swallowing strong magnets is pretty dangerous.


I have a hard time explaining this idea, so please bear with me.

As a founder and hacker, I want to know how my users use my web and mobile apps. But I haven't found an intuitive and dead-simple way to do it.

Mixpanel and google analytic are way more complex than what I have in mind.

What I would like to have is more similar to Facebook, but instead of looking at my friends activity, I'd like to see what's happening with my users.

I like the idea of a feed, or more specifically an exploratory object graph.

I'd like to open my phone and see a feed of what's happening, similar to a cat | tail but where I could click on a verb (Joined the website) or a noun (10 users) and it would direct me to that specific page.

So, for instance, it would tell me:

15 users joined <My service>. Bob And Alice got a new highscore of 2500. Marie played level 3 10 times.

I could change the Zoom ratio.. I.e. see the feed of the last minute or the last year, and it would sum up the actions to one statement.. "1541 users upgrated to premium", etc.

Basically, I don't want to look at graphes or stats or parse log files, I just want a beautiful and simple feed of what's happening. And when I see something that I'm curious about, I can just click and learn more. Just a dead simple way to understand how my users use my apps.


I faced the same problem while I was building a SaaS product. It became difficult to understand what users were doing on the app and then easily reach out to them with help/feedback emails. We realized that there are other people building SaaS products who are facing the same problems. So, we have pivoted to building an analytics + messaging tool for SaaS products.

These are the problems that we are working on:

- quickly check out the feed of any user's actions on your product

- click on a action and see the recent / most-frequent users for that action

- segment users by their 'journey', e.g. case: how many users have integrated their Google Analytics account but have never seen the generated reports? (user 'hasdone' and 'hasnotdone' queries)

- auto segmenting users based on their engagement/frequency/plan (by assigning a karma score to each user)

- lifecycle emails / notifications : send emails to your users based on where they got stuck while using the product etc

You can check out more on the landing page:

http://www.dodatado.com/landing

My email: prateek [AT] dodatado [DOT] com

Links:

Website - http://www.dodatado.com/landing

Beta Sign Up - https://dodatado.wufoo.com/forms/dodatado-account-registrati...

SaaS product I was working on - http://www.dodatado.com


This is the philosophy behind http://spinnakr.com. It is still a work in progress, as getting the thresholds right and making sure events are the important ones across a whole range of different sites is on ongoing effort, but I totally agree this is the way analytics should work. Right now we're just focused on traditional web analytcics but you should see user focused data soon as well.


Have you checked out Intercom.io? It does a lot of what you are asking ...

https://www.intercom.io/


I like it. It is like Mixpanel, the implementation on the page would be similar, but without data aggregation.

There are open-source Mixpanel-like backends out there, some will not come with a frontend full of charts, you should just get one of these and build a frontend full of feeds.


https://github.com/paulasmuth/fnordmetric


Mixpanel has released this exact feature you want now.


You can do that by using Google Analytics custom events. https://developers.google.com/analytics/devguides/collection...


I believe KISSmetrics is what you may be interested in. You send in events, and then are able to see who has done which event, create graphs, and do more powerful analysis.


Seconded. I've used KISSmetrics for this exact use case. There's a bit of groundwork required to setup 'events' but once that's done they will track every user and you can see how the user ended up using your app.


A service that figures out how to create and deliver healthy (meaning not extremely salty, sugary, etc.), same-day-prepared, vegetarian-friendly meals for $10 USD per day or less. The service would make one or two dishes so that they can be prepared in batch (maybe one non-vegetarian, and one vegetarian), and deliver them at a predetermined time. The current services in this industry tend to cost a lot more than this and/or require you to do preparation yourself (i.e., they supply the ingredients only.) Even better: if you can accept the delivery soon after "cook time", you can get your meal warm, assuming it's not a cold dish to start with.

For people who don't, can't, and won't cook, and realize that much restaurant food isn't focused on trying to balance taste and health, this would be a great service. $200/month, 5 days a week. The logistics part seems feasible based on what's current in the marketplace; the economics are an open question.


Not quite the same, but these guys are close: http://www.platejoy.com/

It differs in that they deliver the ingredients, you spend up to 30 minutes cooking (for dinner) or 10 minutes (for lunch).

(no affiliation)


I suspect this would need to be city-wide, or otherwise regional to make sense from a logistical point of view.

But the most likely criticism is the lack of variety. You send somebody food they don't like. e.g. A meat-eater who doesn't like beef, and they'll be unhappy regardless of how healthy it might be. Suddenly you need to let people choose from a menu, and the economies of scale become harder to achieve as you've got to cook N-types of meal, not two.


Does Sprig or Spoonrocket or the other alternatives not deliver healthy food? I believe they have a vegetarian option.


Spoonrocket looks the closest (CA only for now, I'm in NYC), so thanks for that. I'll keep watching them. I couldn't find a subscription option for Sprig; the others mentioned upthread were "ingredients only" services. Looks like this model is getting closer, though.

I think that a company like Spoonrocket experimenting with a subscription service would be worth trying - the companies would have steadier income, and customers wouldn't have to think about it unless they didn't want the daily meal.


Not exactly the same, but when they started I found them to be impressive in delivery speed, Berkeley/Oakland: http://spoonrocket.com/

(I am not affiliated in any way)


Does http://blueapron.com not fit the bill? $10/meal /person I think? Ship to > 85% of the country.


If this would be a Kickstarter project right now, I would put a month worth of money on the table. Low salt, low sugar food you can't get anywhere.


Yup this would be amazing! The closest thing I could think of would just be to pickup/have delivered Native Foods every day.


While we're at it * :

Funding Monday - where everyone pitches to VC's in the same HN thread.

Trial Tuesday - where everyone provides a link to their demo in the same HN thread.

Writing Wednesday - where everyone writes an informative hacker-related topic on their blog and shares a link to it in the same HN thread.

Throwback Thursday - where people all share something informative about past experiences or past apps/etc. in the same HN thread.

Free Work Friday - where people request work to be done by volunteers to develop their product with a possible chance at employment at some future date all in the same HN thread.

* - with slight sarcasm


I like some of these ideas. The Trial Tuesday could be a nice way for people to get feedback on their WIP projects and perhaps motivate them to keep up and have weekly updates.


that's a great idea. get together a community who will each publish an update/demo on their project each week on the same thread. I'd certainly be up for joining in; it might shame me out of letting side-projects sit untouched for a month at a time.


Okay, a bunch of ideas. Probably nothing as good as my last one (https://news.ycombinator.com/item?id=7542610), which was meh to begin with (but I'm happy to see that it sparkled a bit of interest).

#1) An anonymous blogging platform. A site that obviously doesn't log you, and that allow you to publish posts without having them attached to an username. Since many would like to somehow post under the same handle, but without having the post attached to an username, implement a system like tripcodes on 4chan (for anyone that's not familiar, you choose a random string, and then you get an "username" derivated from that string. No passwords, no login. As long as you're the only one that know the original string, you are the only one that can post under that "tripcode"). Problems: Money, it's hard to make profitable an anonymous site. Useless, it's not that hard to pick a random name and create a blog on wordpress.com, blogger or similar while using tor. Spam, a site where one can post anonymously without hassle is just asking for it.

#2) (totally inspired by Idea Sunday, but I've had this idea before for a while) a site to post ideas for a project, or a "I'd totally love if someone where to do this". Then implement a rating system or something similar, so that the most common/interesting ideas are always at the top. And a random button to get a random idea (ideas as service!). Call it "Jar for ideas" or similar :)


#1 http://pen.io/


#2) Something like http://www.reddit.com/r/somebodymakethis ?


I'd like to see something like #1 where not even the site hosting the content can impersonate an anonymous blogger. I suppose this would require some kind of key signature system. The problem with tripcodes is that while they offer a persistent identity across every site that implements the tripcode algorithm, you're giving up your secret password for the trip to every site you post on.


I created something similar to #1: http://millisay.com

It hasn't really taken off, but I also haven't put much effort into promoting it. It still lacks a few features (I've been busy with other projects), but I really like the clean, simple design and a few people have posted some interesting stuff.


Nice :) One of my ideas about "tripcodes" was that eventually an user could upload a style-sheet to customize his blog, one could see all the post by a certain tripcode and could follow him (rss?).

I agree, I like the simple design. One problem (or maybe I'm just dumb): I've pressed the "New" button a bunch of times to post something New, and only after that I realized that I should have pushed "Sumbit" instead. Maybe rename it to "Recent"?


I knew nothing about tripcodes until now, thank you. Won't it be a drag to have to type in a tripcode each time you want to post something? Isn't it better to allow having a username and password, but hash whatever is displayed as an identifier on the screen to preserve anonymity?

I'll consider adding tripcodes to something similar I built. It's a bit of an experiment: http://hikigo.com

On spam, it's not anonymity that leads to spam. You get spam on sites that have username/passwords too, like HN. You can fight it in the usual ways any site fights spam.

(A note for HN moderators: your algorithm for detecting spam links automatically deletes a new post from a new user.)


>I knew nothing about tripcodes until now, thank you. Won't it be a drag to have to type in a tripcode each time you want to post something? Isn't it better to allow having a username and password, but hash whatever is displayed as an identifier on the screen to preserve anonymity?

Yes and no, you could just say that the tripcode is a password, and then the user wouldn't even have to type the username :)

Also you can save it to localstorage or something, so that the user has to type it only once.

>On spam, it's not anonymity that leads to spam. You get spam on sites that have username/passwords too, like HN. You can fight it in the usual ways any site fights spam.

I agree, let me say that differently: such a site would have a very low barrier to entry, easily exploitable by bots. Just type in your post, eventually a name/password/tripcode/however you call it, and press the publish button. Easier that the usual register, confirm valid email, publish.

>(A note for HN moderators: your algorithm for detecting spam links automatically deletes a new post from a new user.)

Are you using tor? Because I remember reading that if you create a new account on tor you have to wait two weeks before your account is allowed to post.


Thanks good feedback, thanks! I've changed it to "Recent". The tripcode idea is very interesting...I'll definitely have to think about implementing it.


I made something like #1 a few years ago, but it never took off and I eventually grew bored of it. It's on Google App Engine and open source if anyone is interested to continue with it.

http://wikiblog.jugglethis.net


Regarding 2) Weekend hacker (http://www.weekendhacker.net) used to be a cool little email list with side projects people needed help with. Maybe expand your idea to something like WH?


for #2 I walways think about building a personal note-taking service for ideas, further developments of ideas and ideas that generate other ideas, in a tree-based structure.

It would work as a personal thing, but it would also do some similarity search and show ideas from other people that are similar to each of your ideas "near" them.

So, if you found the others' ideas useful, you could link the two and proceed from there. The other part would see their idea got linked by you and will come and check your idea, maybe establish contact, maybe not, maybe just ignore the linking and proceed working alone.


For #2 : Try Betterific https://betterific.com/

Does pretty much the same thing, allowing users to post ideas/problems/solutions.


If you want to share ownership with a community of builders you can post up ideas to https://assemblymade.com


IHMO, key to well executed #1 is that it should be social yet anonymous so the writer gets a sense of community without feeling like they would be judged.


A buddy of mine built #1 once - dont think it really took off though...


for #2) check out http://sideprojectors.com ? :) Reddit also has a subreddit called "somebody make this".


FYI your chosen font probably looks beautiful on mobile, but does not render properly on desktop Chrome or Firefox for me.


I had a small discussion about this idea[1] and I believe that it is worthwhile:

Make private key based authentication useful for logging in to websites. I don't have the chops but introducing private key based auth as an extra can be very worthwhile for some users. So start there and improve the UI until your grandmother can use it.

I believe the biggest hurdle will be private key management. The advantage of using private keys for authenticating with web services that if the keys become unrecoverable the keys can actually changed by the service provider (after sufficient diligence).

[1]: https://news.ycombinator.com/item?id=7553882


Check out virtual smart cards, a very cool feature: http://www.windowsecurity.com/articles-tutorials/authenticat...


Steve Gibson is working on a system called "SQRL" that does exactly that, and should be very usable: https://www.grc.com/sqrl/sqrl.htm

edit: forgot some words


You know client certificates can be used for this, right?


Having spend the last week or so working with client certificates, they could be so much more if browsers implemented a proper interface for managing them, or at least a much more intuitive one than they do at the moment. They work spectacularly for building closed APIs where you control both sides, however. They're just a bit tricky for the end user most of the time


Yes, see the linked discussion. I think the hard part is getting people to use them.


A usability test of mobile apps delivered to consumers through ad space on free mobile games.

You create a "use quest" of your mobile app. One or more screenshots of your application with a stated goal - e.g. invite a friend on a social app. You present the user the home screen and he has to find where in your app he can invite his friends. You record all clicks, and time and etc until he finds it.

This test is served as "partners quest" on free mobile games. The gamer has to complete your "use quest" in order to continue playing that pool game. You pay the game producer a small fee and charge from the mobile apps developers that want to test their designs.

You may give the gamers more in-game bonus if he allows you to record his voice and face (and pay the game producer a little bit more for this feautre).


These guys are doing it for a whole app. Maybe get in touch with them. http://techcrunch.com/2014/04/11/app-io-turns-ios-apps-into-...


A subscription service for open source projects.

As a company I would pay you monthly and get access to bug and features development for all the open source dependencies that I am using. The more I pay the more things I can get done in a month. If a feature is too big for my budget I can accumulate month to month value on a ticket or get other companies to participate.


I'm not taking a side (really!), but there's a lot to be said for keeping open source projects strong on their own without injecting cash. Cash coming in creates a disincentive to develop without cash and promotes a reliance on paid services rather than a mentality of contributing back. Also, when it comes to many projects, the organization that offers services around them has a negative incentive to contribute clear, quality code.


> ...there's a lot to be said for keeping open source projects strong on their own without injecting cash.

Money is already all over the open-source world. Many smaller projects will not involve it much, but it matters in the larger ones. E.g., much of the code contributed to the Linux kernel comes from people who write it as part of their jobs. That means that feature additions & bug fixes in Linux will be largely those that matter to (say) Red Hat. This idea offers a way to get work I want done prioritized without needing to know how to do it myself.

It also offers the possibility of software development midway between the start-up model that groups like YC and Kickstarter are pushing, and the no-money-involved-ever model. Surely somewhere between the two we might find something workable.

> Also, when it comes to many projects, the organization that offers services around them has a negative incentive to contribute clear, quality code.

A good point. I don't have an answer for that one.


> Also, when it comes to many projects, the organization that offers services around them has a negative incentive to contribute clear, quality code.

I'm probably being naive, but in my version of this idea, I was thinking that contributing well-written, tested, and documented code back upstream from the company's account would be a good way to have the work that's done contribute to the marketing.

Part of the reason why I haven't moved past just talking about this idea is the possibility of it having a negative impact in terms of the original creators/maintainers, so I'm definitely interested in hearing more about this perspective on it.


I believe this is called a RHEL subscription... in seriousness though this is a good idea and wouldn't be that hard to put together. You would monetize it by taking a cut of the spend and passing the rest on to the projects.


This is a lot like gittip and BountySource.

Furthermore, a project maintainer could put on his README, that, for a certain recurring donation, you get some level of direct personal contact.


This sounds interesting, but doesn't make much sense. Do you mind explaining a bit more?


It's a way to shift the value proposition. Nobody in their right mind is going to pay for free things. You can try to appeal to people's good will in funding rounds but it requires a recurring effort that detracts from the work. This is because we still have the view of software as being a thing you get in a box. A specific delivery that you use for a while until the new version is on the market. This is how proprietary software like Microsoft Office is structured.

Now in practice, software is never a done deal. You have bugs. Platforms evolve and you need to adapt the software to it. New features might be added. This is all things you might want to see happen in the future. A subscription service could be a way to participate in the process indirectly. Obviously things need to be balanced as to not corrupt the projects in question.

There are already actors in the place like RedHat and Canonical or product companies like MongoDB but they are focused on their platforms and usually sell "support". As a customer I have a big and changing variety of open source project upon on I depend and negotiating contracts with each of those is time consuming. I would rather pay a fixed amount and have access to a body of experts and direct (part of) their work on projects they own or not. If I scale my subscription up I get more throughput.

A lot of details are conveniently left open to keep the narrative simple and idealistic :)


Interesting to see this here, I had this idea a while back and have been discussing it with people before I go on validating/invalidating it.

In my mind anyway, the idea is something like https://railslts.com/ but for potentially every other gem in your gemfile. We would provide security patches, bugfixes, features, documentation, training, version upgrade scripts/documentation and so on to subscribers.

I really like the accumulated credit for features, maybe a system of like: request feature, we put a "credit" value on the complexity of the feature, then companies/people can pledge credits/money for it would work?


I've also had this on my mind for quite some time.

How would you actually implement things though? After all, you're essentially trying to sell "support contracts" for open source projects via the platform. You could implement these by a) "hiring" the project owner to work on specific bugs/features or b) hiring freelancers to fix requested bugs.

Furthermore, how would you ...

* determine the value of bugs/features? * react to a project owner abandoning a project? * handle multiple upstream developers?


I hadn't really cracked how to scale the business up, but to bootstrap it I was thinking that I would just do the work until it became too overwhelming. I haven't thought about what pricing would look like, so I don't know how feasible it would be to just hire more people as more and more work comes in - but it should have some interesting economies of scale since there is a lot of overlap in the gems used in rails projects for instance.


One important part as a provider would be to have a way to collect all the dependencies on which a customer depends on. This helps you priorities your maintenance work if the customer doesn't have any requests (auto-pilot mode). Is is also useful as way to sell: "see all the things you depend on". And finally it lets you take decisions on which project maintainer to poach.


is this what you are meaning? https://deps.io/


This is a good start. Once you have all the customer's dependencies it becomes easier to offer value-added services like maintenance and feature development.


I like this.


I want a batch notification service. Every 15/30/60/90/configurable minutes (or on request), the service would ping all my communication channels and give me a summary on a page or to a mobile app. I don't have notifications for email/social media because I don't want the interruption but, during downtime, I find myself checking through all the different services I use to see what's outstanding. I don't need the app to interact with these services, just ping via API and return what's "unread."

I tried to use IFTTT to build something workable but the key is getting notifications all at once for everything when I want them, not duplicated notifications when they come in.

I'm willing and able to work on this with someone if anyone wants this problem solved (PHP, JS).


I wrote something to bundle "alerts". It was a node application which would receive updates via HTTP-posts.

From there they would be injected into a redis queue, and fanned out to email (daily), a simple HTTP web-page, and xmpp.

It wasn't hard to setup the core. The hard part was pushing notifications into it. I suspect you might have the same problem. Writing a deamon to stream twitter and POST items matching a regexp, etc. Good luck integrating with facebook too!

Most of the pieces are out there, and joining them up isn't hard. But .. it didn't work out as well as expected just because of the integration aspects. That said I have a semi-working system pinging me via XMPP when new blogs are posted, etc.


I don't know that batching the alerts would solve the problem, I want to be able to ping these services to determine unread counts at that point in time. I'm, of course, just assuming that these are available via API from the services I use.


Blackbox web security SAAS.

There are some services out there that will scan you for vulnerabilities, but I have yet to find a really slick way to integrate security "for free" into your own servers/services.

This would mean at minimum:

1. apt-get install X. Automatically decent kernel/network/firewall configuration occurs, removes cleartext passwords from ssh and changes the port, etc etc. This is just a base starting line.

2. malicious user detector. Probably cooperating with an API within ruby/jvm/python/php/whatever is a mixpanel-like way I can define custom events and possible milestone-firewalls. E.g. this is a new user sign up, here are the relevant fields & ip. Or here's a comment on an image from user X with IP Y. Rather than having to figure out if we're likely being abused outsource that decision to the service & have it help you block the evil-doer.

3. APIs to help with usernames/passwords, setting up TLS & perfect forward secrecy, etc. That is to say, give me patterns I can use that every app has even with some level of customization.

4. Have built in DDOS protection, particularly for resource starvation attacks that might be more app-specific (request X took Y>Z time, and it's being repeated endlessly).

I understand that security is complicated, ideally should have full time staff, every app/service is different, ya yada. Except when you're starting out you don't have these kinds of resources and something like this could get one quite far, not to mention be able to see much of the internet and what's happening across clients (like gmail with spam) and thus have an advantage over you. It seems like a giant hole in the world of app development and most people are terrible at security when it comes down to doing anything more than firewall + ssh keys + bcrypt user passwords.


So this would be similar to Impermium (now acquired by Google) meets Sift Science meets Cloud Flare - have you checked them out?

I have thought of building such a security as a service for websites for sometime. My PhD thesis was on building a reverse proxy server (kind of like nginx) to detect attackers (via machine learning) and starve them of CPU/network resources disproportionately (using admission control + WFQ) during a DDoS attack. After that I've spent a decade building systems to detect cyber security threats, click fraud, Twitter spammers, etc.

What do you think is lacking in existing security as a service offerings that prompted your post? Would love to chat more.


Social Researching: Like social bookmarking, combined with joining comments so you can share and collaborate on your research.

Think "trails" from V Bush's "As we may think" article, implemented so we can collaborate. As users start bookmarking things in their own trail, the app works out what other trails are similar and suggests you merge your researching efforts.


This idea has a /lot/ of use cases. I'm not sure how easy it would be to make a business out of it -- maybe you allow public collaborations free, and charge to take it private.

It'd sink or swim on the details of the UI, though, and I have frankly no idea how to get those right. Lots of painstaking user testing, presumably.


Google wave was great for almost this, actually, minus the matching / merging part! Sad to see it go.


Yes. I think this idea is almost in many things but you need to give the user the structure to both read and contribute to a thread of research.


A pay-to-post/view social discussion site.

Reddit starts out cool, then slowly becomes 'dumber' (discussion-focused articles turn into quick reaction images) as it becomes more and more popular.

Hacker News avoids this by purposefully not trying to be as popular as Reddit -- UI remains minimal, tries to stay out of search engines' ways, etc.

But, in this way, HN is discussion focused through obscurity. What if you filtered out dumb crap by making it a cost to the user for posting? In that way, the user would have to really care about what h/she's posting about before h/she posts it. I.E.: thought-provoking discussion and not cheap images.

That way, you can mass market and accept huge gains in popularity while being sure that the people who are posting/voting on the stuff care enough to pay for it. Want cheap (often not even funny) images? Go to Reddit.

The best posters get compensated with the money that the voters paid. This creates a network of "I pay to post and I pay to vote but I get money from posting which incentivizes me to post but I only get paid if I get voted on a lot which means I'll want to encourage voting which means I'll have to vote a lot myself" : the voters and content contributors conflate in one genius network of money movement.

It still wouldn't ever get as big as Reddit, simply because people don't like direct payment for services on the Internet. But it'd be an effective way to get a bunch of smart, dedicated people who would kill for thought-provoking discussion on the Internet, and also pay them for just that.


The best posters get compensated with the money that the voters paid.

Any service that offers a significant monetary incentive to cheat it, will implode.


Adsense is pretty far from imploding.


Money is not the solution. Reputation is.

People are not currently penalized for contradicting themselves, being dishonest, being thoughtless. But they soon will.

Let's cure the web's Alzheimer.


Exciting rhetoric.

How do you expect people to discern between Reddit reputation and TheWebsAlzheimersCure.com's (drop the "the" -- just WebsAlzheimersCure. Cleaner.) reputation?


How do you handle users gaming the system for profit?

I was thinking up various ways to introduce money to our communication systems to incentivize certain behaviors. But not from the modern Facetwitter perspective of a site that wastes your life by addicting you. The true measure should be the correctness of the content and money should pay for that.

First a system where you have to pay to post and if you post is confirmed as not spam or abuse, then you get the money back. Otherwise the site takes your money. Should be a negligible amount. Mostly a way to make spam expensive and compensate for moderation, but of course a site can abuse a user this way as well.

If a user posts some profound story/comment and others want to compensate him directly with an easy UI, they can add a monetary amount to their upvote. But there is little guarantee that the amount will be significant or the money will arrive promptly. Users should be able to undo money sent for well sounding but ultimately wrong info. This is to discourage participation driven by points and monetary rewards.

If participation and content creation is driven by money even more than it is now, there will be a few new legitimate sources with useful information but the vast majority will be content farms. To discourage this "points" or "money" should accumulate, BUT should not be paid out for years until the accuracy of this content is confirmed by the passage of time, peer review, etc. If mistakes are found, all readers should be notified of corrections. There should be penalties for releasing misleading or mislabeled information (like tech blog posts that are really advertising but pretend not to be. You should pay to advertise and pay more for advertising what turns out to be crap after peer review.)

A site should take its share for moderating content well and working on rules that incentivize useful content/services and penalize attention seeking of little utility.

It's hard to predict effects of various new incentives. But a site should work on modifying the rules until desired results are reached. Users should not stray from normal behavior to satisfy some quick and dirty addicting point systems. If that means limiting participation to select trained users, releasing information to larger audiences only after it has been discussed by relevant experts (so no shocking attention grabbing studies are released) then so be it.


I like this idea a lot, thank you for sharing it. I might try this with http://hikigo.com. Like most things it might be hard to get the initial core of users.

What I liked about the parent post is: It still wouldn't ever get as big as Reddit. When you find yourself thinking that way, you are probably onto something you actually want.


That's the thing. If this service is implemented properly, you want users to game the system for profit. It'd have to be so airtight that the only way to "game" the system would be to go out and try to think of something intelligent to contribute. Another commenter discussed the possibility of astroturfing -- The astroturfing would have to be so well concealed that it'd be content users would actually like. So, it's still corporate-supported advertising, but it's smart corporate-supported advertising. Which is still awesome. Reminds me of this : https://xkcd.com/810/.

What you said about a system of experts and trained users determining whether your comment is worth exposing to the public ... eh. It sounds like it'd take a shit-ton of time and debating, especially as the service expands to handle a lot of content.

Also, isn't that just modern journalism? As a journalist, you apply to write an article for, or get a job at writing articles for, a well-established high-quality media company like the NY times or the New Yorker. A system of trained experts review your application and approve or deny it accordingly. Then you write articles and they get shown to the public after you've been approved.

I'm not saying it's a bad process, I'm saying it already exists. The problem is, it's slow and it takes a shit ton of time to determine who's a great writer and who gets limited participation and who doesn't. And when you allow the process to get faster, you have to give up the expert-authority, and well, you get something like Reddit.

This service would have to find a way to keep submissions and distribution extremely fast, like Reddit, but also keep the quality of discourse high, like NY Times.

The proposed solution is : money. People who submit stupid cat pictures to Reddit aren't dumb themselves, they've just been exposed to a forum for discourse that allows dumb content. Meaning, if you asked them : "Do you really think that cat picture is going to allow for quality discourse?", they'd definitely say "No."

So I'm willing to trust ilovekatpictures1998 on his judgment of discourse, if not his ability to create such discourse. So if we can trust the users to have good judgment, all we have to do is create a system that forces them to use their judgment, which Reddit doesn't do. And telling them to put their money where their mouth is by making them pay is the best way to force them to use their judgment.

The only reason it wouldn't get as big as Reddit is, people don't want to put their money where their mouth is. Well, most people. But then again, there are a lot of niches ...


The social discussion site MetaFilter uses part of this idea already. It requires users to pay $5 once before they are allowed to post anything (http://www.metafilter.com/newuser.mefi). But it does not charge or pay its users after that.


I can easily see it becoming a platform for astroturfing and viral marketing campaigns by corporate accounts, and i'm not entirely certain i'm comfortable with directly linking a poster's ability to participate in a discussion with their disposable income.

Would be an interesting experiment to see though.


The astroturfing would have to be so well concealing that it'd be content users would actually like. So, it's still corporate-supported advertising, but it's smart corporate-supported advertising.

Reminds me of this : https://xkcd.com/810/.


That depends on the degree to which money equals power in that setup, I suppose. It doesn't have to be smart if you're willing to pay for enough accounts just to create a buzz.

Although granted - for the site to be worth the effort it would already have to be popular.


Instead of cash, use a cryptocurrency. Dogecoin fits perfectly with this.

As an aside, I'd like to see a torrent site that uses a cryptocurrency model to control access/leeching/improve seeding. I'd rather fry an ethernet card than a GPU!


Isn't The Well pay-to-post (at least in a sense since there's a subscription fee)?

http://www.well.com/


I wish I knew. This site seems more interested in telling me about it's wondrous history than giving me a forum of discussion.

Can I get a frontpage? Links to different subforums? A design that isnt from the late 90s?


Email Link: A simple browser plugin which allows me to click it once and it emails me the current url. It would be nice if it gives me the option to email it to others.

Resuscitate Xoopit, a much much better gmail search. Have you ever wondered why gmail search sucks when its parent company is the supposedly the best search engine in the world. Xoopit was a nifty addon that mine your inbox and 1) organize every file and links it found, and 2) gave you a search feature at least 10 times better than the gmail option.

I would pay once for the first tool, and monthly for the second product.


I just wrote a rudimentary version of “Email Link”. You can use it for free. You just need a few lines of JS in a bookmarklet:

  var myEmail = "my-email@example.com";
  var url = window.location.href;
  var subject = "Link: " + document.title;
  window.location = "mailto:"+encodeURIComponent(myEmail)+"?subject="+encodeURIComponent(subject)+"&body="+encodeURIComponent(url);
To install it, change the email address in the code, URL-escape the code (with `encodeURIComponent(…)` in a JS console), prepend “javascript:”, and save it as the URL of a browser bookmark with the title “Email Link to Self”. Put the bookmark in your bookmarks bar or somewhere else convenient.

When you want to use it, open that bookmark to open a new message in your mail client, then click Send without changing anything and switch back to your browser. To send the link to others, edit the To field in your mail client before clicking Send.

P.S. Firefox, Chrome, and Safari all have a menu item for this built-in, except they don’t prefill the To field with your email address for you. The menu item is in the File menu, about 3/4 of the way down.


"Email Link" : Something you want to save for reading later? Pocket and other services do the exact same thing, that is, they save links for reading later, and you can email it to yourself too.

In case you want to automate the process, you can create an IFTTT recipe for it. :)

And you don't need to pay for it.


pocket even provides an rss feed of saved items so you can read and manage them in your reader.

No affiliation to pocket and no idea if other similar products do the same thing.


Yes! There are other apps too, like Instapaper, Readability, etc. Pocket looks good, and it has browser extensions so you can save the current page with just one click. The app is one of those most beautiful apps I have seen on Play/App store.


I've been wanting this but with a little more functionality. The Email Link with a dropdown of up to 5 people I email the most, so that I can select one of them or more and email them something (the link of the page I'm on or additionally a select text from the page I'm on). No opening Gmail, no creating a new email, no subject line, no cutting and pasting of links/text, no hit send.


OneTab kinda does the first.


A site to organize systems.

There are plenty of "List of frontend development tools" or "list of git tools".

I want a site where users can build lists, share them, vote on them, categorize the content and update it with the new tools.

alternativeto.net is a partial attempt at doing that but it doesn't take a systems thinking approach


I had exactly this in mind last summer and actually built something! http://curate.im - check it out. I had planned to add voting and collaboration functionality, but nobody ended up using the site so I gave up on it and it's just been sitting there for almost a year. It's not tailored specifically to systems organisation though. Enter the beta code HN001 if you want to sign up lol


I've always wanted something like this to. Only, I suspect that lists are only one possible form of information that needs to be organized. I also think users should be able to shape trees (and maybe others).

Maybe you encourage vistors to click on everything they agree with and thereby learn which users contributions are seen to be the most valuable. Users with the most valuable contributions could be rewarded with more ability to shape the tree or list...


List would supported nested lists i.e it would be a tree

In terms of rewards or monetization, it would be possible to gamify contributions ala stackexchange and rely on affiliate links for commercials tools and perhaps even share the revenues with the top contributors or offer them options on what to do with the money


Maybe something like http://pineapple.io/ ?


http://getthinglist.com


Sounds like squidoo, but I don't think anyone uses them anymore.


somewhat similar: http://learni.st/


Fix podcasts.

I was helping someone set up her android phone for podcasts and noticed how problematic the whole process is. First, the android apps aren't very good. More problematic is the process of getting a feed into it. Sometimes searching by name works. Often it doesn't. Getting from podcast's website to the app is a problem. Discovery is also quite poor.


I am using http://www.shiftyjelly.com/pocketcasts. That is really amazing. And the discovery have not failed me yet.


Seconded. Pocketcasts is amazing, especially the settings. For example I have have mine set to keep only latest 2 podcasts pershow, auto download on wifi, and auto delete those I've listened to. It's a well built app worth way above $3.99. Don't bother buying other podcast apps or even show apps like This American Life.


Yup.

I switched from iOS 6 months ago and podcasting was one of the initial problems I had. I trialed a handful of apps before properly purchasing pocketcasts. As good as the iOS app.


Plus, the devs are awesome and have fantastic release notes. :)


A couple if German podcasters are about to fix this: the have crowd sourced an antenna pod fork to make it a single purpose app. Toby Bayer is the main organizer. Alle the code is open source: https://github.com/tobybaier/Einschlafen-Podcast-App


Latest commit by tobybaier about 2 years ago


On iOS it's not much better, most apps are too granualar (e.g. Downcast) or not enough (Apple Podcasts). Pegging some hope on http://overcast.fm/ fixing the situation but it's not a great ecosystem...yet.


The built in iOS podcast app is the worst app I've ever used. It's not that it fails to be great, it actually seemed to go out of its way to delete podcasts just before I left home. I bought downcast and I'm happy with it except for the problem of discovering podcasts.


Most free podcast android apps are pretty terrible, but PocketCast (paid) is really awesome (at least it have a great UI). I don't even listen to podcast that often, but I bought it without hesitation.

But it's true that there's a problem with podcast, so take my +1.


I've been using BeyondPod for a couple years now. It has a paid and free one and the free one has suited my needs, but its a relatively cheap upgrade for some nice features.


Throwing in a mention for PodKicker[1], since you mentioned searching/discovery. The UI needs a lot of work, but it is nice to have the search engine built-in (they claim an index of 250,000), instead of trying to grab RSS links from a podcast's homepage. The search itself is even available as a separate app[2].

[1]: https://play.google.com/store/apps/details?id=com.podkicker

[2]: https://play.google.com/store/apps/details?id=ait.podsearch


That's the one I use and helped her get started with.


I work on an iOS app called AGOGO that takes a different approach than most podcasting apps. Instead of subscribing to shows you browse channels podcasts, tv clips or text to speech audio. Would be interested to hear what you guys think www.agogo.com


I nearly worked on an idea like this 4 years ago. End-to-end podcasting which helps people make good quality, listenable podcasts. +1 from me


Maybe just the idea of podcasts is broken. RSS has dying a slow death for quite some time. Why not just build a better podcast?


When an HN post is popular, we sometimes crash small websites/blogs.

How about a URL Shortener that caches the page at the time the shortened URL is created?

Then, when the short url is used, the service attempts to reach the site, if it fails, automatically inform user & redirect them to cached page.

Or, the service can attempt to reach the page periodically (once/ couple minutes), if fails, redirect all users to cached page until next attempt in a couple minutes. This period can be based on number of users clicking on the url. Once the link is no longer very popular, the service will have a fairly long time between attempts.


The free Coral CDN (http://www.coralcdn.org/) already does something like this. It caches the link the first time it is visited. But rather than redirecting you to the original page if it’s up, it always show its own mirror of the page (on its nyud.net domain).


Something like http://archive.is ?


This is the idea behind http://fireballed.org/. Nice idea though.


Like a link shortener built on top of the Wayback Machine?


Blockchain Genius. Crowd-sourced annotations of transactions in the Bitcoin blockchain. Clever UI for 'zooming' in and out of display of the chain.


A crowd-sourced (á la Wikipedia) AI platform to create a chatbot (á la Siri), with a nice user-friendly front-end, so users can input de "inputs and answers".

I guess that with time it could be improved with APIs and better algorithms.


This is a great idea!

I'd be terrified of the results if /b/ get hold of it. I saw the results of a few chatbots and it was funny, but also scary.


Yes. It's been tried but people kept giving it useless data.


I'm really interested in chatbots, though it seems the underlying state of the art is still pretty poor.

Mitsuku (http://www.mitsuku.com/) won the Loebner Prize last year, and that chatbot still runs on basic AIML files.

I think a new approach is probably necessary to take chatbots to the next level, but I don't know what it is.


That's creative enough to have potential to win the game of AI.


MOBA sports. That means DOTA or League of Legends style Soccer, Football, baseball, any team sport, really.

Baseball probably wouldn't work, but soccer would be fabulous.

- 5 on 5 with 10 minute halves, every player on the field is a real person.

- Pay to unlock famous players and skins

- Just bring FIFA controls over to a multiplayer game.

- Players can form their own teams and play on ladders.

- Start it in Brazil, then translate to other languages.


You can do this in FIFA already (since FIFA 11 I believe) - it's 11 vs 11 multiplayer. You start in a lobby and then pick your position.


Do people actually play this mode? Is it popular?


"Who's in town" app - it's difficult to keep track of where all my friends are at any given time, and many times I miss out on catching up with friends because I don't realize they're in the same city that I'm traveling to.

One could say something like the "location" field in chat programs deals with this, but hardly anyone keeps that current. In the past, I would blast a Facebook status of my new location and hope that people would notice - this sort of worked, but it was by no means perfect (and I have since deactivated my account there)

Essentially, you'd specify a whitelist of people you're willing to share your location schedule with, and the app would alert you if there's overlap.


Check out Connect. They launched recently and I think they resolve what you're hoping for:

https://www.youtube.com/watch?v=pVKzyoGbm-c


Or, would it be possible to extract location information from existing Instagram/Facebook/etc friends via their posts?


Google Navigation for Cooking. So as Google Maps gives you turn by turn instructions for driving, an app could give you line by line instructions for cooking. So often when following a recipe, I have to read the next item on the instruction list for a recipe. I think it would be great if I could have these read out as a line - by - line item that I (or whoever i'm cooking with) can then follow. I love cooking with people, but sometimes having to stop conversation mid-sentence to read what we next have to do is cumbersome/awkward, and having an auto-pilot for new recipes that would read the next "task" for cooking would be awesome.


Dark Ages: randomly shuts down all functionality of your phone, forcing human contact, at random 10-20 minute intervals , no more than a couple times a day.

Perhaps there is a machine learning aspect that learns when you are being anti social (ie playing Candy Crush for 3 hours straight) and is more likely to "dark age" you.


Tracking browsing habits. I would like to see how much time I spent on what "kind or type" of browsing. And I can define these as a hierarchy (type->subtype). And define productive/unproductive/lessproductive segments.

Example: "stackoverflow or localhost:8081" would be "Development" kind of work. "news.ycombinator.com" --> "Reading" "gmail.com/outlook.com" --> "Commmunication" "facebook.com" -- "Unnecessary waste of time" ...and so on.

The reports would simply show- how much time I spent on what segment (Productive work/ General Browsing).

A chrome extension/app would be a good starting point - personally I do not use other browsers unless absolutely required.

Further I would also like it to automatically classify and show what kind of content I spend time browsing - like "Entrepreneurship / Programming / Nuclear fusion / Space Travel / ... "

If anybody wants to try it - let me know. I can dedicate a very small amount of time on this.


You might want to check out BeLimitless (http://belimitless.co/): It's a Chrome extension that uses your browsing activity to tell you how productive you are/ have been.


Have you tried RescueTime?


Thanks, looks promising - I have seen couple of other desktop time-tracking apps and this looks much better than those.


Social graph app for tracking gang activity on Twitter, with the intention of predicting potential violence before it happens on the streets. According to Vice/other sources, it's becoming increasingly popular for gangs to communicate publicly via Twitter and other social media sources. It's an interesting NLP problem because Twitter language is crazy and an interesting "computational social science" problem because it could actually help communities. You could partner with (or sell it to) municipal police departments.


I was listening to NPR a few months ago and they actually had a bit on tracking Twitter for gang activity and the ethics behind scanning twitter. Here's a link to the story [1]. One of the products they outline is called BlueJay [2] and scans for key words and locations. I guess NPR isn't popular among gang members or they would probably stop.

[1] http://www.npr.org/blogs/alltechconsidered/2014/02/28/284131...

[2] http://brightplanet.com/bluejay/


thanks!


oh cool, pre-crime. worked for tom cruise, why not poor people in disadvantaged communities?


:(


Haha, I think your idea is great, I just think it has unintended profiling consequences and is the sort of thing that justifies police overreach. Even compstat (crime tracking system that largely helped clean up NYC and then other US urban centers) is being intensely scrutinized for its role in driving profiling, harassment, and quota arrests in places that aren't war zones. The police aren't great at understanding how systems like this can be used to make people feel safer rather than feel oppressed.


Yeah... I totally agree that this is a huge risk, and in fact I would expect abuse to happen because cops. Also, to be honest, I really don't like the idea of making software for the man. However, I'll probably still work on the idea this summer, just because it'll probably be an interesting project :)


Offline navigation map for mobile. Small thread here:

https://news.ycombinator.com/item?id=7171919

The entire OpenStreetMap dataset is 23 GiB with moderate compression. With some optimization/truncuation the whole world should fit comfortably on most phones.

This is very useful, for the times you're lost with a slow connection, no connection, in a foreign country, or with roaming charges.

Also it's a fun algorithm-heavy challenge even if you don't finish anything.


While offline navigation is increasingly an edge case I do agree that it offers vast utility in the situations where it's still necessary.

Nokia's HERE Drive+ offers this as its headlining feature on Windows Phone, and Windows Phone comes with offline mapping support built in (to be downloaded on a region-by-region basis).

I think the most challenging part of offline mapping is honestly geocoding. Decent vector data for streets is easily available thanks to OSM and learning how to render it and do routing is a fun path down learning well-known algorithms. When you think about it, you're basically reimplementing early 2000s standalone navigation systems like TomTom.

But even once I've implemented rendering and routing (and reached early 2000s TomTom parity), I still can't enter "Pizza Hut" and end up at the nearest Pizza Hut. This is where Google excels over other map providers and especially over the existing mobile-phone offline maps solutions like Nokia's - Google's PoI database and search facility is unmatched in my experience.

Google even solves for fun problems like locating the closest instance of an intersection between common street names like "23rd and Broadway."


Google maps will cache the current map view at the highest detail resolution if you type "ok maps" into the search bar and hit go. You don't even need an active SIM to use gps and zooming. I cached all of central London and used it from my otherwise non-functional Verizon CDMA iphone for months.


Unfortunately, Google maps caching works as view-only. There's no offline navigation at all, I imagine that's all done serverside. Only the tiles (or now vectors) are downloaded.


That's not completely true--if you load your route on wifi and start turn by turn navigation, you can be completely offline.


Do you mean navigation as in "routing" or just as in "viewable map"? If the latter, MapsWithMe does a good job of it already.


Um, never mind then.

http://wiki.openstreetmap.org/wiki/MapsWithMe

12 GB, impressive.


>for the times you're lost with a slow connection, no connection, in a foreign country, or with roaming charges.

I have had exactly this problem this week, so it's been at the top of my mind.

Algorithm-heavy is an understatement. Tasks that are more naturally done by the server (geocoding, routing) have to be shifted onto the client.

Bear in mind that the cases you describe probably are also limited-battery and limited-memory.

There is also some existing competition in this space that you would have to catch up to. Since UI acts as the main differentiator to users in my opinion, that probably ought to be your primary concern at first.

Having said all of that, I would be super-excited to be involved with an Open Source/copyleft project like this.


MapsWithMe[0] lacks routing, but it is an iOS app that provides fast, detailed offline maps with POI's. Data is downloaded by region (for the US, it's by state, for most of the rest of the world, it's by country) and the data is updated pretty frequently. When I was trekking in Nepal, even obscure trails were marked on the maps, which was terrific!

[0] http://mapswith.me/en/home


Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: