I've been a little quiet lately as I've been working on the next demo before the end of the Kickstarter. Based on the reactions I've gotten so far, I think you guys are in for a treat :)
What's most exciting about this whole thing, is the prospect of a new and potentially, hopefully, more flexible mode of sustenance for software developers world wide.
The initial financial risk of the develop first, market later model seems to be greatly mitigated by the kickstarter approach. The market can be validated in real terms, with real money, before a developer has to take the plunge. Striking out on your own was (perhaps) never so easy.
The new model may very well liberate many developers from the financial shackles that bind them to their day jobs; it may create more stories of the Marco Arment lineage: independent, flexible, balanced.
With each new successful story like Light Table that comes along, the more convinced others will become that independent software development as a living may very well be possible for themselves as well.
I advise caution - successful funding does not yet mean successful product. The greatest threat to Kick-starter is the strong potential for projects to devolve into marketing drives for funding followed by people running off with the money, because they have no liability.
One of the big budget endeavours will fail to come to fruition, and the entire internet will throw a gasket, and then we will see if the model is sustainable without having contractual obligations from the fundee.
I can attest to that. My project met it's goal of $25k over a year ago when the largest project was still < $1M. It was great market validation. But I was surprised at how much market validation was still left. It didn't deter me though since I know building a business is hard regardless.
I think $200k is still low enough that the market and product needs to be vetted. Pebble at $10M isn't such. Only they can screw it up themselves at this point.
I'm excited that the goal is reached. One thought: I just saw your "We made it" blogpost via planet.clojure.in. May i suggest that you post it as an update on kickstarter?
I think most people on kickstarter expect and enjoy when the project author "celebrates" the reached goal with them. In addition the news about the 2nd demo is remarkable.
Btw any forecasts as to when the demo is going to go public?
Looking forward to it. I didn't contribute at the initial rates, so was delighted when you took on board the comments here and rethought the pricing (if that's the right word).
I'm complete ignorant at programming languages. But I've seen this video a few times, and it still seem like each feature are things I already do every single day of my life as an Android developer. Live documentation, organizing your work around functions, watching values and expressions change on debug etc. These are all things I already do every day.
So suppose lighttable adds support for Android development. How exactly would lighttable help me? What exactly can I do with it that I cannot do right now?
Are you guys backing this up just because it looks visually different? Is it because the languages it supports just don't have good tools for like Java does? Or am I missing something obvious here?
I see here a possible cyclic thing happening with development environments. I've watched a lot of friends abandon IDEs in favor of more lean setup like textmate, or going the direction of emacs.
Lighttable strikes me as a possible move back towards a full IDE environment. A lot of the things shown here are already available on Eclipse but it's presented here with a much cleaner visual design.
The one thing Eclipse doesn't do is the "live" debugging with value substituting. I do think that would be very, very helpful. I could see that being implemented as an Eclipse plugin as well and it would be extremely useful in certain cases.
I hope LightTable has really good keyboard based controls. The reason I prefer terminal environments more is just being able to not have to constantly mode switch between keyboarding and mouse pointer.
It seems to be based on http://codemirror.net/ so you could probably check that out and see the keyboard support. I wonder if the final product will be a browser-based editor as well..?
You sound dead on to me. Java tooling is basically light years ahead of JavaScript, just because it's so much easier to write for a static language. Most people don't learn all the advanced features of Eclipse and other IDEs and debugging until they are stuck writing in them for years and years anyway. I know how to do things in hot keys in Eclipse that I don't even know how to do otherwise. Hot code replace and debugging and all the other stuff is similarly an advanced topic, even for many Java coders, so it is no surprise JavaScript programmers think it is something new.
I've done development in Eclipse off and on for a few years. Some Python (PyDev extensions), some Java (Android, school assignments, a desktop app). I haven't tried it all that recently, but one of my biggest complaints has been performance. I end up turning off a significant number of the features that people talk about for Eclipse, simply because they make the environment run too slowly. It's not as if I'm using underpowered machines either...
I don't want to talk for Chris Granger, but my intuition tells me that the features you listed are things he considers to be the baseline for a modern editor and are not the true reason he's building Light Table. I think his goals are more about making a real product with the capabilities of the demo IDE Bret Victor showed in his Inventing on Principle talk https://vimeo.com/36579366 Eclipse is nowhere near this today, and perhaps it never will be. I'm backing Light Table because I think Chris has a shot of getting there.
Hmm. I work with Eclipse a lot, but this is not just visual. Maybe I use something different but the way of organizing your methods on screen and the actual live coding I don't have in Eclipse. We're not talking 'debug' here; we're talking live coding. Did you check videos like: http://www.youtube.com/watch?v=7XUWpze_A_s If you can do that with Eclipse & Android, please tell me how because then I have been stupid (which is very possible: I haven't changed my toolchain in a while).
That's the first video I found on that stream. Note that it's about 50 seconds in that he starts actually editing the code; he edits the slide friction before going up the slope. There's a better one of notch designing a dungeon game for some competition where he's writing the rendering code and it's changing as he does it.
That's eclipse right?
You can do the same in visual studio as well, just need to have 'edit and continue' enabled. It's one of these magic things that you either know about or don't, but generally if you're a web developer you'll never have needed it.
I'm not sure what you mean by We're not talking 'debug' here; as technically that's in debug mode when it's running at the same time as coding.
Functionally it seems no different to me. Notice how the 'draw player' function in the clojure example actually contains the variable for color, it's not a pre-initialized variable. I wonder whether storing and changing that elsewhere would have resulted in a colour change (logically I would guess no otherwise the game state would contantly reset).
Cool. Yeah I use that and it's interesting how I didn't connect that to light table at all. Must be the bling, but yes, it's the same thing. I guess it's because, like you said, I usually do web dev in Java and don't really need it. In Javascript on the other hand, it would be very welcome (Firebug like stuff but then in IDE). Thanks for the insight. Light table is still an interesting experiment besides anything else.
I like light table and all that, seems a good idea as IDEs are too heavy but for me:
1. I don't need the docs all the time as I've already got that given I use a statically typed language 80% of the time and IDEs actually already do that
2. Most of that other stuff is already available in Eclipse and Visual Studio
So for me it's mostly shrugs what are you all so impressed by?
the part where variables turn into literal values for a specific test case - that's something I haven't seen an editor do - is that a mode for IDEs that I don't know about?
I just imagined all the time I'll save by never stepping through with a debugger again.
If anybody else is interested, here is a Web Numbr graph of the funding for Light Table over time: http://webnumbr.com/light-table-pledged (Sadly, I didn't think of setting this up until it was really far along).
Light Table is an interesting approach. I think, it would turn out to be something different & interesting if the project leverages that aspect of it. There are lots of editors which have already done a fine job of editing code. Unless a new interface to computers comes along, those are here to stay.
What I am trying to say I guess is that, if the project separated the approach and editing aspects of coding, and then integrated the approach aspect with already existing editors (or an API/protocol of sorts, like what X server is to the various WM/DE); I believe it will turn into something really interesting.
I'm inclined to agree. You cannot beat Vim and Emacs as editors... the key is to play a game they can't. I mean to show some examples of that with the next demo :)
Maybe I'm misreading your reply, but I think you might have misunderstood gits1225's comment. What (s)he was saying is that it would be great if your system would be open so that it could (eventually) be integrated into existing editors. Your reply, on the other hand, talks about "beating" these editors--to me that doesn't sound like you're interested in interoperability.
For what it's worth, I wholeheartedly agree with gits1225. I think will be very hard for you and your team to even remotely approach the editing facilities offered by Emacs or Vim, seeing that these facilities have evolved over decades--especially in a web editor which is subject to a number of constraints.
If your system could somehow (again, eventually) augment these editors--or, if you prefer, be augmented by them--rather than attempt to replace them, then in my mind this would be a win-win for everybody.
Yesterday, as a last resort after a day and a half of unproductive frustration, I printed out a bunch of code from five or six different (too tightly coupled!) files that I was trying to refactor, spread them around my desk to get some sort of handle on it, and attacked them with a red pen.
I guess my concern is that that's pretty far down the page and since the the fact that the $200K goal has been met is prominent, that message might not get across very well.
This is a really interesting method of funding a software project. It doesn't seem sustainable though. People might (maybe) be willing to fund a "Light Table 2.0" project but certainly nobody would care to fund an "ongoing maintenance and bugfixes for Light Table 1.0" project.
Perhaps the Kickstarter model could be adapted to fund maintenance and bugfixes too. Popular open source projects always have the problem that their bug trackers get spammed with passionate complaints from well-meaning but frustrated users. I'm imagining a public bug tracker where users can pay to raise the priority of their favorite bugs. Not necessarily a "bug bounty" system, but just a way of raising money for the project overall while simultaneously giving frustrated users a productive outlet and reducing spam.
I may have missed something, but I believe the point is to get it developed with the help of Kickstarter-money in the first place and then sell through the normal channels to fund future development. As incentive for backers, they most likely get away with a cheaper price than the one we'll see later when the first version is out for sale.
It's not entirely clear how it can require a license and be open source at the same time. I suppose we'll see how open it really is when it's released. But selling licenses doesn't preclude also having a way to pay for bug fixing priority.
I know someone who wrote GPL software and sold alternate non-GPL licenses - apparently (at the time, at least) some organizations had very paranoid lawyers who won't allow any GPL software to be used. He called it the "Sap" license. Those particular companies tended to have deep pockets, so this worked out.
some organizations had very paranoid lawyers who won't allow any GPL software to be used.
I was working for such an organization a few years ago. On top of that, one of the founders would even post "anti-communist" rants against Open Source on the company intranet.
Being open-source and selling doesn't have to exclude each other. You just have to trust the buyers moral, as in, if they're using it to create things they get paid for, they also pay for a license of your application.
Also, since when has having a closed source ever prevented piracy anyway?
That's a really big if. If it's truly open source then there will also be multiple contributors. Who do you pay? People always say this "open source doesn't mean you're not allowed to sell it!" and that's technically correct, but I don't think this strategy has ever been demonstrated to be economically feasible.
Then again Red Hat's value add is the support. Red Hat's business model is not to rely on its customers' altruism. There are various reasons to dislike such a model, especially for an IDE, not the least of which is that the point of a well-designed IDE like LightTable is that it's easy to use and does not require a support contract. Being a support company creates the conflicting incentive that the software should not be too easy to use. Furthermore, support is not fun; you need to be a big company to earn a significant amount of money, because revenue is directly tied to man hours. Making new and exciting IDEs is fun, and revenue can be more directly tied to the value you produce, rather than the hours you put in.
There are various reasons to dislike such a model, especially for an IDE, not the least of which is that the point of a well-designed IDE like LightTable is that it's easy to use and does not require a support contract.
Lots of BigCo are paranoid about the support contract thing. Sometimes partners and investors want to see support contracts in place as a security blanket. In those cases, the support contract for a "well designed" IDE in your scenario is just free money.
JetBrains seems to do alright selling IDEs. In their case, I think people pay for the value, and they continue to pay for upgrades. Support is through the web, so it fits in with revenue commensurate with value.
Yes, indeed. JetBrains uses a closed source model and that works really well (they have a "lite" open source version, but the version they earn money from is closed source). I'm not saying that you shouldn't do support at all, just that it shouldn't be your primary business model.
As I understand it: the source will be freely available for you to download, modify and compile. But if you want some nice convenient binaries then there'll be a small charge.
Since Light Table relies on CodeMirror (AFAIK), will some amount of money go to Marijn Haverbeke, along with other kinds of contributions to his project?
This IDE/text editor has some very good ideas, but I would never use it because it doesn't seem like it will be open source, free software, or even freeware.
I wouldn't rely on a text editor that can't be improved and changed by the community.
Besides, it seems to run inside the browser, and most browsers (even for a single tab) consume quite some RAM.
