Hacker News new | past | comments | ask | show | jobs | submit login
Still running strong after 1,532 days without a code change. (ryandetzel.com)
361 points by ryandetzel on July 13, 2012 | hide | past | favorite | 130 comments

The first contract program I wrote was when I was 17 years old. I was working at a service station changing oil. Some guy comes through who wants to stand around while I do the work (it happens) so we struck up a conversation.

Turns out he was a bookkeeper and had just purchased an Apple IIe and wanted to use it for his clients. I knew nothing about accounting, he knew nothing about computers, so it seemed like a good match :)

Four weeks of spending free afternoons at his shop, and it was ready to go. He was happy and I had 200 bucks in my pocket. Life was good.

Almost 20 years later, I get a call from him. He says the program isn't working so well and he wants to upgrade. I'm like WTF? Does anybody in the universe still even have a working Apple II anymore? Why would he keep using something like that for 20 years?

He told me that as computers modernized, it became a bit of a status symbol to have an older-looking system spewing out reams of reports. His customers, who were mostly small construction companies and such, got the feeling of stability and security from something that was unchanged.

It is a very strange feeling to get a call about code you wrote a long, long time ago. If I would have had any sense, I would have realized from the experience that programming is normally an extremely tiny part of actually making a business work. But it took me many more years to figure that one out.

I have a similar story. Fresh out of college (in a non-computing related discipline) I went to work as a temp for a gas pipelining company. Their data entry system was obscenely awful- an Access database that was used like an Excel spreadsheet- entering in the same data over and over again in different forms. So, I optimised it. Added some forms. Made report templates that printed out automatically. Eventually I stretched the limits of Access and turned it into a full VB.NET (shudder) application. I tried to get the IT department involved but they refused to have anything to do with it- I couldn't even install Visual Studio Express because I didn't have admin rights on my machine- I had to use SharpDevelop.

All this happened while I was still an "administrative assistant" on a temp wage. After a time, I left- it was time, and in any case they'd said that my work was going to be replaced by an integrated SAP solution that would span the entire company's workflow. I moved on, got a much better job and have been programming off the back of that original job ever since.

This was in 2006. I spoke to a former co-worker for the first time in years about six months ago, who told me to my jaw dropping surprise that they were still running my application. Aside from anything else, I was blown away that the thing still worked. There's a certain morbid curiosity that makes me want to look at the code I wrote back then, I'm sure I would cringe. But hey, it's still going.

There's nothing as permanent as a temporary solution. :-D

Stewart Brand's book How Buildings Learn: What Happens After They're Built has a great chapter on MIT's Building 20.

> Building 20 was a temporary wooden structure hastily erected during World War II on the central campus of the Massachusetts Institute of Technology. Radiation Laboratory member said, "At one time, more than 20 percent of the physicists in the United States (including nine Nobel Prize winners) had worked in that building".

> Due to Building 20's origins as a temporary structure, researchers and other occupants felt free to modify their environment at will. As described by MIT professor Paul Penfield, "Its 'temporary nature' permitted its occupants to abuse it in ways that would not be tolerated in a permanent building. If you wanted to run a wire from one lab to another, you didn't ask anybody's permission — you just got out a screwdriver and poked a hole through the wall".



this building gets mentioned in a lot of management books and "where good ideas come from" to illustrate a good collaborative environment. it has been replicated by quite a few places with various degrees of success. its main function is that there are no permanent work places so that people are forced to collaborate. This has not always worked out well.

could it be that MIT happened to have a lot of smart people and they created good things while working in building 2.0, and they would have created just as many great things if they worked in another building?

maybe im just extra skeptical because im currently reading "Fooled by Randomness" :)

It's quite possible that the building was special and did have an effect that was lost when people tried to replicate it. Trying to replicate things is usually a losing game.

For example, I bet they didn't replicate its temporariness or the fact that you could make holes in it.

Good point. It can be easy to conflate cause, effect, and coincidence.

The lesson I take from the story is that the smart people had extreme freedom to pursue their research. I don't think a university can just create a dumpy building and hope to produce some Nobel prizes. <:)

This is one of the biggest things I'm grateful my mentor drilled into me: Never make a quick and dirty solution that anyone else (especially management!) sees -- it will become the final solution.

You know, the great thing about learning is that everything you learn turns out wrong. And after that the next thing will be that you learn that it's wrong that it's wrong. And so on.

So guess what, you will probably learn next that quick and dirty is an amazing thing. Sometimes the morbid, buggy stuff is exactly what's needed and a well designed solution wouldn't scale to it's task or wouldn't be used because it's so clean.

This story (and others on this page) should drive home the point about how little most businesses really care about technology. If it solves a problem, they do not care about the hardware, programming language, database, user interface, etc. Look how many business applications still use character-mode interfaces (most banking, insurance, stock/inventory systems, etc). Why? They work.

This is exactly how I got my start in software. I was hired to walk a warehouse at a furniture retailer and take inventory with pencil and paper. At this time I had 0 computer experience.

Soon I was voluntarily transferring the paper to spreadsheets and printing them out. Next was a starbase db with a simple form on a laptop.

A year later I had built a custom system with a mysql database, barcode scanners at the arrival and departure gates and automatic email reports to the owner.

It is because of this that I was able to get to move on to another company to be hired as a software developer without a college degree.

Ditto. I did some temp work for a Big Dumb Company that was trying to manage compensation on a gigantic Borland Quattro Pro spreadsheet. I sold them on converting it to Paradox, bought a book on programming in Paradox and haven't looked back since.

I urge you to get a look at the code and see just HOW cringeworthy it is. For science!

Ha, well it was written in VB.NET, so that alone ought to make me cringe. And I was building on my existing knowledge of VBA, so...

Apparently a few years ago they had a minor panic because they realised they had no idea where the database was actually stored. I'd gotten some MSSQL space from the (outsourced) DB guys- the guy who gave me the space had also, of course, left. Fun times...

i've seen some great software built in vb.net including a real time trading platforms capable of near high frequency execution speeds connected to 10 exchanges simultaneously.

i personally don't care for vb.net because of its notation, but i don't think it's THAT bad. you're making it sound like it was written in vb6

I wrote a manufacturing pull system in VB 6.0 + Access for the application database. That was all they had to work with there ( or more correctly all they wanted to work with.. ) It handled a stupid number of queries per month and was very reliable.

It's important to have good tools, but a robust design beats all imho.

Loved your story. And because it's what we do here, I'll mention my best similar war story. I created a bowling alley scoring system in, I think, 1990. It was basically an embedded system, using commodity PCs (one PC per two lanes) networked together with a master PC running another one of my programs as a check-in and control console.

I used Turbo C V2 I think it was, with my own home rolled DOS windowing library of the kind that was ubiquitous at the time. Basically it was a smell-of-an-oily-rag development for a friend of mine who had taken over a near derelict 20 lane bowling alley and was upgrading it from "manual" scoring to "computer" scoring but couldn't afford a 100K "automatic" scoring system from Brunswick or the other big player (can't remember their name).

Anyway, it was a successful and fun project, but after 8 years or so the bowling alley was gutted by fire. Years later I was travelling through a small provincial town and saw a 10 pin bowling alley. Something told me to stop the car and go in. What do you know, there was my whole system, lanes plus controller, running a small 8 lane setup. Talking to the guy (potentially a mistake, what if he wanted after sales service :- ) it transpired he'd bought the equipment in the (literal) fire sale and figured out how to hook up the network and get it all going. He'd even split it in two and got it running another system in another nearby town. I was seriously impressed that he'd managed that actually.

> Something told me to stop the car and go in.

I am a rationalist and have no truck in telepathy, crystals, or other such related craziness. Coincidence and the law of large numbers explains everything.

...but once in a blue moon I hear a story like this and do have a niggling thought: maybe the universe is much deeper and weirder than we think. And/or the grad students who are running the simulation we all exist in like to play tricks on us from time to time. ;-)

I am also a rationalist and similarly have no truck etc. But this was a rather weird experience, in the sense that I clearly remember some inkling that I should check out this bowling alley, for some unexplained reason. And no, I have no particular interest in bowling and would not normally go near a bowling alley.

Law of large numbers.

We also don't know how many other bowling alleys he's gone into - he presumably likes bowling.

I don't disagree with you, but I would note that I don't really like bowling as such (I rolled a few games when I was working on the system, but never really got any good or enjoyed it that much). So no, I don't make a habit of going into bowling alleys.

I know someone who went to work for a small company in the early 70s and among other things wrote a cron job in PL/1.

She came back to the same company 30 years later in a much more senior position. And out of curiosity checked to see if any of her old code was still there. She found that cron job she had written in a language that she no longer knew and found that it had never been touched. She asked why not, and was told it never broke.

Unfortunately that company imploded in the mortgage crisis, so her code is probably no longer running.

That's gotta be a good feeling. Knowing something you wrote has worked flawlessly for 3 decades, even if it is something trivial.

My uncle works at helicopter factory nearby as technician, part of his job is to maintain old database that keep temperature readings from meters inside some oven for making plastic parts (hull, etc).

The system works on old PC with MS-DOS, and database is in clipper. I think the system works for at least 15 years already.

Actually the lesson to learn is charge monthly subscriptions.

Actually that's a little unfair. The thing is these are businesses where (presumably) vital parts are left to fester. I presume because the management 'dont get it', but every business has these corner cases, and odd parts. I think the foolishness of ERP was ever daring to dream it could be solved monolithically.

THe future surely is a form of SOA where some guy in accounting can write a crappy service, authenticate it globally and plug it into the enterprise directory system.

But this really assumes everyone working will be able to program. Just like we assume everyone at work can read.

Stories like these make me think that the whole software industry is a huge conspiracy theory :)

Thanks for sharing

My story:

Whilst on a university placement doing some web coding and databases work, the guy in the office next to me was having real problems with his website.

Being curious, I asked for the URL and took a look. A classic ASP and Access database powered website that was probably held together by bubblegum. He had outsourced the work to a Chinese company because they were cheaper than UK based companies (important point for later). The site was full of every web security hole known. Mainly lots and lots of SQL injection holes. He asked if I could fix the pile of garbage and I honestly couldn't, I don't like rewriting but when its an old (this was 3 years ago) tech with no English comments and their code quality zero, copy pasting everything.

So in the evenings and weekends for a month I rewrote it to ASP.NET MVC (1.0, latest at the time) and SQL Server. I got him to select a template off Themeforest for $15 that I reworked slightly (and badly, I'm not a designer).

He is happy, I am happy with £500 for my work. I finish my placement and go back to university but came back to my placement for another year to do a startup idea. He had moved offices and we have very little contact.

After finishing university and 2 years later, I join a "Soul Destroying Corp" (not the one you are thinking of) and decide that isn't what I want to do with my life, so I setup my business and tell him that I'm now doing my own thing and looking for clients.

Turns out my weekend/evening £500 website generated him around £500,000+ in revenue (not profit) and his business grew from just him to 5 people. All his sales leads came through the website.

From September last year till now, I have rewritten it and we are automating more of his "manual" (lots of Word, Excel and paper) to be totally automated or a few mouse clicks.

Wow, when I think back, that website was terrible but it worked well for 2 years and grew his business. Huge ROI for him but now he is paying me back by hiring me to rewrite/automate more of his business.

How did he manage to find you after 20 years?

Phone number portability.

Looked up his name in the phone book, I assume.

Upvote for the story and for the last paragraph realization. Unfortunately many in our industry still don't understand that.

Your story immediately conjured up an image of Steve Martin in The Jerk.



This was a joke. :(

This is just great. I had a very similar experience with http://openexchangerates.org - hacked it together and left it running, and only realised just how popular it had become when, one morning 6 months later, I woke up to over 20 emails, all saying similar things to what the woman said to OP.

"What the hell! It's been down for a day already! When is it coming back?"

I was pretty bemused, so I checked the server logs and found that it had been hurtling along at what looked like 60,000 requests per day (I later found out it was closer to 200,000).

Needless to say I jumped in, fixed it, and got talking to the users who relied on the project for their businesses or personal projects - it's been a blast.

edit - it's since been completely rewritten, so I can't say the code went unchanged..

ha! i love open exchange rates! - i hit it once a day with a side project of my own. thanks for running a quality service!

Thanks! Glad to hear it - email me if you'd like the side project to go in the showcase.

> I set a reminder(thanks Siri) to check these stats later that night.

Thank you, alarm clock! Thank you, spoon! Thank you, hammer!

In any case, thank you for the post, and thank you even more for keeping this kind of service running for so long :)

If you've never thanked your car, you've never lived anywhere that it gets cold.

If you still start your car when it's cold, you've never lived anywhere that gets really cold.

(Starting your car at -55C is not a good idea, even when it has >300W of heaters keeping it warm.)

Where the (freezing) hell? Not even in the Uruguayan Antarctic base does it get to -55C (my personal worst has been -17C, and I've been to Canada and the Austrian Alps).

Or been on the thruway when you're really, really low on gas.

How advanced does an AI have to be before you should say 'thank you'?

It actually frustrates me that Siri doesn't give a hoot whether you thank her.

I want to thank her a) because it feels natural and b) to tell her that she's actually got something right. She's usually less use than an ice kettle so I want us both to get a kick out of it when she's not.

It could actually have function value as it could trigger some positiveness in a bayesian filter, subsequently increasing results accuracy. I let you be creative about the negative sentences.

Perhaps they'll add a thanks-awareness feature that will eavesdrop on your online conversations and thank you next time yyou're using it - "oh BTW, thanks for mentioning me to your friends on your blog" ... or y'know, maybe not.

The key for the everyday user probably won't be better AI, but rather the achievement of more natural interaction rhythms. People cajole, beg and bargain with their cars, computers, baseball teams; with no response expected, the rhythms are conversational (if wry).

The one second press-and-hold, then the three second wait for response, are obstacles that keep Siri removed. Once those limitations are inevitably gone, and one can immediately say "thanks" and hear a naturalistic response in less than half a second (like Malkovich in the commercial where "time sequences are shortened"), the illusion will be a strong one. Statefulness, ubiquity, and responsiveness will be a powerful UX combination. To the degree Siri interaction becomes naturalistic, we'll probably say "please" and "thanks" much as we do in everyday speech.

Seems like those changes will just cause us to enter a stage where we'll hit the uncanny valley. What Siri seems to overcome is some of the social awkwardness around voice interaction, but the last mile, so to speak, will be much more expensive to overcome.

Hope I'm wrong though.

once it gets moody if you don't?

I got into college on an essay addressing essentially that topic. It was a while ago and discussed what emotions are appropriate when a talking ATM thanks you for banking. In a British accent.

Apparently he showers with Siri. So naughty.

The nice takeaway from it is that the code you wrote works perfectly well for a large number of people and is obviously stable and bug-free enough for them to use it for their businesses. Code isn't about hammering out new features or moving fast and breaking things.

And if it's written in Perl looking at the source code again might break it so I wouldn't do that.

Is Perl a quantic programming language ?

Well, how else do you think http://www.bennylingbling.com/tag/schrodingers-bug/ originated

My professor's favorite joke:

Java: Write once, run anywhere

Perl: Write once, run away

My best example is a Perl bug I corrected.

It occurred only 27 second by year! Just to think about the probability to detect it knowing the code is used only about twice a week.

I'll bite: how did that happen?

I don't remember all details. My co-workers searched to understand the bug and he discovered that changing the creation date for just some second made the bug disappear. I then discovered he called the function with a bad argument. Strangely using an integer instead of a time object (I am not sure about that but you get the idea) will render the exact same result most of the time. I corrected the bug by using the right function call. I then tried a brute force search to discover for how many time the bug occurs. And if remember well there were two close period of a few second each during wich the function bad called render a different result. It was something like 17 second and 3 minutes later another 10 second. I believe I simply tried for the current year. I might have tweeted about it some years ago.

Zombie projects are fun. I created this in 2006 and haven't changed any of the code since, other than the copyright date (in 2010 - I guess I should update that).


I didn't add Google Analytics until 2008 but here are the stats since then:

    Visits: 304,619
    Unique Visitors: 246,368
    Pageviews: 361,097
    Pages / Visit: 1.19
    Avg. Visit Duration: 00:00:41
    Bounce Rate: 84.44%
    % New Visits: 80.87%
I get one or two thank yous a year from people that get my email from my whois entry.

I'm not sure you have to update the copyright every year for any particular reason. Anyone know?

I know, and you don't have to. (Actually, you don't even need the copyright notice at all -- copyrights are valid without it -- but it's a good idea anyway.)

In fact, you shouldn't update it. If you write something in 2002, someone steals the code in 2005, and in 2007 you change the copyright to read "copyright 2007", then later it becomes an issue, the person who stole it can claim "Look, my version is earlier!" and you'll have to go through the explanation of what happened and perhaps provide proof of it, which is an argument you wouldn't have had to make if you'd just left the copyright notice alone.

I guess I just use it as a proxy to show how long the site has been up. I use a range in the footer on that project:

© Copright 2006-2010

Probably dumb but I do find myself looking in the footer of sites to see how active it is.

We do this, but make the later date dynamic. So it will always say Copyright <year>-<current year>

Is that legal? If you don't make any changes the copyright date shouldn't change.

Since those notices aren't required, I don't know whether they have any legal status at all.

Theoretically, the copyright date tells you when the copyright will expire: a work Copyright 2000 will become public domain ten years sooner than one Copyright 2010. However, since neither of those seem likely to ever expire in the first place, this mostly only matters with very old works (around the start of last century).

Good article and a good reminder of how what is just a side project to the creator can be so important to its users.

It looks as if he's updated his forum software about as regularly as the site's main code because it's full of spam (http://forum.invoicejournal.com/). Might be worth fixing that as the consistent number of sign-ups and invoices suggests there is a community of users there to be fostered.

10 years ago I wrote a sales system in Perl with a couple of friends. We charged something like US$750 (it was really simple: uses a hash instead of a database). It is still going strong (no bugs reported) and about US$500.000 has been sold with it. That's what I call ROI and stability!

Totally off topic, but I really dislike how ruby (and I guess others?) has popularized the word hash to mean a hash table. I stare at that sentence for a good two minutes thinking "how does someone use md5/sha etc instead of a database???" before I realized what was meant. There are plenty of good (and well known) unambiguous names for this data structure that I really don't understand why it needs to be given a name that's in common use to refer to something different (and only kind-of related really). Some alternative names that have been in common use since the dawn of time are hash table, hash map, map, dictionary or associative array.

I think it's the Perlers who started saying that; it predates Ruby's popularity. (And I agree, I still stumble over it every time I read it.)

hash is know as nosql in modern world.


Ops. Sorry 'bout that.

After reading the entire article I noticed it doesn't contain a single link to the actual project. I like your style sir.

Would you believe me if I told you that was a conscious decision? :-)

If he's already nervous about old code, why send a huge traffic spike there...

This is the living proof that the "web app market" is really stratified and the value perception within each customer group varies immensely. Who would tell that a 4 year old scope of features could keep a product alive? Who could possible agree that a system that supports thousands of accounts could survive without "hygienical" processes like OS/DB maintenance, bug fixing etc.? Who could imagine a product with a growing adoption/acceptance not having paid options included as plans to leverage revenue in the long run? Uff, this is all mind bogging to me. Congrats to the author though, this is an incredible story!

Gives me a bit of a "ghost in the machine" feeling here. You created something interesting on a whim, gave it life, and sent it off into the world. Unbeknownst to you, it chugged on, somehow successfully urging people to use it. It did this mostly on its own.

I'm sure there's a white paper or talk on the life developers give their software...

What keeps hitting me over and over is I created something very basic, very quickly for myself and now years later people actually rely on it to run their business. I felt so guilty getting that call...I felt as if I had let this women down even though I had never met her. The site was probably down for a couple of days...how many other people did I let down? How many people lost money or sleep because I failed them?

You might want to consider setting up some basic monitoring. If you don't have much time to put into this, you get one free watcher with Pingdom (no affiliation other than that they watch my website for me), and you'll get an email if the site goes down again.

There is a lesson in this for many programmers. Cool features and complexity are not what's important. What is really important is that the user can do their job.

What's trivial for a software developer can be life changing for a non-technical user.

I can sympathize, I feel this way whenever something I've built for a large audience doesn't work. I feel as if I let them down.

With that being said, perhaps you could make it easy for these users to let you know when it doesn't work. Set up multiple easy lines of communication (twitter, sms through google voice, email, etc..). Find a good, free, monitoring service.

Based on what you told us in the article, this service has been cruising without issue for quite some time. That's better uptime than most projects ;-)

I think you can walk away with a clean conscience if you offer an export mechanism.

Even though I've been programming for 20 years and it seems very mundane now, I am still regularly sometimes struck by the amazement when I'm reminded about how much code I still have out there running, and how much actual work it continues to get done despite the world changing around it and the now-obvious-to-me engineering deficiencies. Digital computers and software are truly an amazing achievement that transcend humanity's inherent ability to create systems.

Please don't sell it or break it down. If it isn't costly to maintain, then don't sell it for scrap as most people seem to want it. Close sign-ups if anything, but free users look like they're happy with it and after trying it, it works nicely. It's something easy to clone and it's not really something anyone would develop on top of, so rather than selling it for its user base (selling your users) plus the concept, open-source it. If anyone wants to copy the concept, it's easy to (and anyone buying it would need to write it from scratch if they want to make money out of it, anyway).

That gives me mixed feelings…

What if the buyer wants to keep the free users happy and make it a profitable business? One thing leads to another, everyone is happy.

If there is one thing that you do need to do, its turn off the support forum. Its full to the brim with spam, and is (probably) the only part of the site that actually does need continuous attention. Otherwise, congrats!

Agile software development must be a conspiracy theory. When a corporate IT manager keeps hearing "embrace change..." whisper all around him, he starts thinking that there must be something wrong with their piece of software that's been running just fine for years. I guess it's good for economy though.

I went into a Whole Foods near my house that was built a few years ago. In the checkout line I looked over and could see their 'terminals' in the customer service section. There were 2 and they we're both running some text window app on XP.

Naturally I thought, that OS came out 10 years ago, has been end-of-lifed, this is a fairly new store, and the eol software is connecting to some text-based something somewhere - who knows how old that part is?

I have a friend that worked for an auto supply-something company that would remote into their customers computers. He would tell me (this was about 3 years ago) that he couldn't believe some of the ancient stuff he would come across on those systems. Tons were Windows95, lots of them had malware running on them, and the business owners somehow kept their business going.

I'm not talking so much about old programs that just run, but more about how businesses can keep old stuff hanging around. Even with malware and such on it. It's obviously just another tool to keep the cash flow coming in. Can't blame them, but such a huge contrast to my day to day which is usually working on/with the latest of everything.

Could have been XP Embedded. Microsoft has ended support, but is still selling licenses for it.

Business needs change. New suppliers with new requirements come in, it has to be adjusted to this accounting change, that personnel change, and Maggie retired, and she knew all the ins and outs of the system better than the current programmers.

That said, if we all went back to keyboard interfaces, green on black screens, and people who used the same software for 10 years and got used to its "quirks", software development would be a lot easier. ;)

Look at the PCs in your bank or an airline, most of them are jut screen scraping a mainframe app that thinks it's talking to a bunch of IBM3270 terminals

What I was saying is that we could build software more quickly if we kept to 3270 interfaces and relied on extensive user training. I know that many moving companies are probably still using software developed for an AS400, and manufacturing still has old software lying around.

We just wouldn't start a new project like that. Software is taking the same amount of time because we keep expecting more from our software.

>mainframe app that thinks it's talking to a bunch of IBM3270 terminals...

...and runs in the simulator.

My bank upgraded from the mainframe though about 5 years ago. This was a painful experiences for everyone, including customers. And now they generate account numbers longer than 6 digits, and customers have to remember longer numbers, appalling.

I can personally* confirm that that is exactly how sites like Orbitz and Cheaptickets worked 4+ years ago. Lots of Java and Linux up front. But deep in the back a bunch of 3270 terminal scrapers to mainframes. But that was then. Hopefully they've upgraded those things by now. Despite the principle of "if it ain't broke, don't fix it" there's also the issue of decreasing compatibility and decreasing availability of folks who know how those things work and can maintain them.

(* If you've bought air tickets or made car rentals on either of those sites, at least back during that era, it's almost guaranteed my code touched your transaction at various points in the search and booking lifecycle.)

From my experience, people will use any free service or product, even if you explicitly say "this product is abandoned".

That's the story with the calendar app I use, http://30boxes.com/

I don't think a new feature has been added in about 5 years and it barely seems maintained, there have been a few day-longish episodes of downtime but other than that it works great.

It has all the features I need and am used to, and I don't have to worry about the product constantly changing. In a way it's almost better to use a neglected product.

Awesome app!

I wrote a site on contract when I was 15 years old, exactly 14 years ago this summer. That site is STILL running. It was filemaker pro + Lasso on a mac server of some sort. http://mactalent.com I don't even want to know what kind of things are wrong with it today, but it made the guy who I built it for many thousands of dollars over the years.

Maybe you should keep it, not touch the code, and consider a new mission to do some marketing experiments with your 2 hours on the train every day... who knows what might happen

While most people are looking for the next {x}-illion dollar opportunity (me included), the positive charge one gets when hearing that something they've created is actually useful to someone else is a great feeling.

Such a great story, makes me want to go fulfill those umpteen other ideas I've got hanging around to see what sticks. (I keep all these ridiculous domain names as kind of a skunkworks project task list.)

I think that is a great counterpoint to the story about the problems single founders have running 24/7 web apps.

It isn't clear though- is this a paid service?

No, 100% free since day one.

Inspired by this post, I went to the website of a company I worked for back in 1999-2000. The login screen for the first "real" web app I ever wrote looks exactly like I left it.

The company was in a bind with some contractors who were developing a web app for them, and I had recently written my own personal HTML site. So, I bought "Teach Yourself ASP in 24 Hours", read it over the weekend, and put together a little prototype, to see if I could actually connect to the database and report the information they wanted. I ended up writing that website for them over the next two weeks. Once I deployed it (no IIS web admins at the company), I went back to my regular Clipper work.

I wouldn't be surprised if it hasn't been touched in years.

Part of me wants to be happy that it's lasted so long. But I'm also a bit disappointed that all the other 'cool' stuff I've worked on since then probably hasn't had one-tenth the use of that site.

> Still running strong after 1,532 days without a code change.

Oddly, I think that's a great definition of "systems software."

Hi Ryan, this sounds interesting. It sounds like something that could use some TLC and maintenance.

If you haven't committed to a decision with it, give me an email (it's in my HN profile).

regards, Paul

From the blog post: "Now I’m wondering if I had continued to work on it, if I hadn’t neglected it could it actually be a useful successful, dare I say profitable project?"

I'd argue it's likely successful just for the mere fact that he _hasn't_ touched it in four years. Had he actually tried to monetize the project, he probably would have just gotten in the way and screwed something up. This gives evidence to the "simple is more" rule.

Thanks for post! I found this very inspiring including some of the comments, since it shows that solving a real need almost builds a business itself with grateful consumers. It also does not require the latest and flashiest software tools. I wish there was a blog that reported these little small projects that are basic and solve a real need without facebook logins flying around and asking me to import my address book.

Hey Ryan,

Id be very interested in taking this over and looking after it properly. Please shoot me an email and maybe we can talk about it.

Thanks Tam Denholm contact@tamdenholm.com

"I had forgotten about it, I had neglected it, I had basically written it off but this soft voice reminded me that all though I had written it off she had not and that she actually relied on it to run her business."

Is this likely to become more of an issue as people move to Web based applications for things they used Office and file folders for before?

I have 2 products in a similar-ish situation, both of which I don't even remember where the source repo is to make a bugfix even if I had to....

But after a few outages that I had no clue about until days had passed, I did learn to sign up for Pingdom so I can at least reboot apache within an hour of the site going down

The site is back up. Google has a cache now in case it goes down again: https://encrypted.google.com/#q=cache:http://ryandetzel.com/...

Heh, one of the first web apps I wrote was around 2002 and it ran on an NT4 box without a modification or a reboot until this year.

I left the company 1.5 years ago and got a call from a supervisor asking me how to start it up again after a power outage and UPS failure.

website: http://www.invoicejournal.com/

It's actually not a bad template. To the point.

Oh, dear, the sign-up button says "SIgn Up" instead of "Sign Up". I'm worried that that is the first thing I noticed.

It makes me a little sad that it hasn't made him any money. I imagine it might be good for the resume, though.

Hey Ryan - I would love to help you with Invoice Journal. Contact me at Hacker @ (My username) dot Org

the website is very slow.. maybe because loading.. consider moving it to a larger server, and monetize it in somehow to pay off!

Yeah, my little Linode box isn't setup to handle over 300 simultaneously users. :-( Normally it's nice and speedy.

* apt-get install varnish * Set it to cache the specific page * Boom, 100k users at once

I was in the process of doing that as you typed that out. I never thought it would be needed though!

Varnish is amazing, just install it by default, it can't hurt.

New post, How varnish saved my ass when I was #1 on Hacker news. :-)

you will definitely want to cache the post, my HN posts that have staying in the top 10 spots for a few hours easily get 600 simultaneous so you can probably expect the same. good luck!

I think both sites have just gone down. :)

Now I remember why I don't use apache any longer. :-)

Well, the HN effect probably isn't helping either...

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