HN Explorer

Ask HN: What do people do with the code from failed startups?

46 points by bobbywilson0 ↗ HN
I see a good amount of failed startup stories come across HN and I always wonder what they do with the code after they have called it quits. I haven't really seen a case of someone open-sourcing their code and uploading on github or something. I know that this is easier said than done, there are probably some legal issues around it. Nonetheless, I am just curious in what people actually do. Do they destroy it? Archive it? Print it out and burn it?

27 comments

[ 3.1 ms ] story [ 68.0 ms ] thread
MrMike
I typically tar it up and store it somewhere safe. On occasion, I'll work on a new project that could use a piece of code I know I've got stored away. Other times, it's something I stumble upon years later and laugh at myself for my coding style & quality, which continues to improve. "Omg I can't believe I tried to solve X that way... wow..."

I've seen a few people open source their code for failed projects. Hadn't really thought about that, but it's something I'll keep in mind should I have an interesting, failed project.

bobbywilson0
Yeah, I am not sure what the real value in open-sourcing the code would be because you can imagine how hard it would be trying to dig through a project that isn't designed like a library for re-use.
jahmed
One persons trash. Functional but failed projects could be interesting learning material.
MrMike
As jahmed said, it could be used as a learning experience. In other cases, it was a valiant (or not so) attempt at making something great, that didn't work out for a variety of reasons. Depending on the scenario, it might be something that I'd like someone else to pick up from and continue, to fulfill the "I want to use this." itch, but I don't have the time/resources to execute properly.
jonah
How do the IP issues work out if you use code from your old startup at a new one? If the old legal entity remains, it still owns the code, if it's gone, then what happens? Are there clauses or things that can release the orphan code to a particular person or company or does it just become public domain since the owner is no longer? How does that work?
MrMike
Certainly something to be conscious of. I should have made clear that for the most part, the code I'm talking about is from entities owned by myself and possibly one other person, a business partner I've worked with on multiple projects for many years. No outside investment or shareholders. This greatly simplifies the IP issues.
jonah
Thanks for the clarification, but legally you'll still need to assign the IP to your new entity?

By default, what happens to IP when the owning company ceases to exist?

MrMike
Oops.. replied to this one level up. Sorry. In terms of the original company ceasing to exist, I'm not sure. Before dissolving, I do the Asset Purchase Agreement.
MrMike
IANAL, but I use a pretty basic Asset Purchase Agreement between the old entity and the new to cover the IP. Again, since what I'm talking about is myself (and possibly one other person) as the shareholder, it's quite painless.
brackin
If the startup isn't costing you a lot it may be best it leave it running. It'll give you more credibility as if someone says "I Made X and failed" yet there's just an expired domain at X you've basically scrubbed out the past and your work.

There's other sides to this as you won't want an ugly broken site up connected to your name. Case by case basis really.

There isn't a whole lot you could do with the code, i'd just say for most open source it and leave it quietly running, if feasible.

MrMike
It's important to note that it's not just about the financial cost of keeping a service running. For example, I had one service that I built over several months. It got some attention, but ultimately fizzled out. It comfortably lived on some EC2 instances I already had running, sharing resources for the project I had shifted into. However, the support requests came in from the handful of users, and while I wasn't able to make it a financial success, it would've been more of a disservice to those folks to provide a partially functioning, unreliable service rather than give them the chance to migrate away to a competitor that could better serve their needs.
p4wnc6
It is donated to starving orphans.
azal
OpenSource it or sell it on Flippa. I sold a previous failed startup code on Craigslist once, it had a website and all. Usually if its just code ppl are hesitant to buy it, put it up with a nice looking site and some pics of what it does and etc, and you should be able to find a buyer. A Friend of mine sold his online business he wasn't making any money in, on BizBen.
speby
Kind of depends on what code you're talking about. Some "failed" startups may have valuable code that can be sold or licensed out, even if it ends being somewhat of a firesale. Others may tuck it away on a backup or project folder and never revisit it or only pluck away some useful bits here or there. Others may use it as a template for a completely different business just to "get an app up and running" if what they had prior had some similar pieces.

Finally, some may decide to open source some or all of the code. As was mentioned, it's hard for others to use it, though, since it is rarely the case that a startup's codebase is going to be pretty and easy to get going and using, versus some nicely structure, re-usable OSS library.

joshfraser
My company EventVue failed. We tried to sell the code, but didn't get any serious offers. We'd raised 500k in funding so any money we gained from a sale would have gone straight to repaying our investors. I would have loved to have been able to do that just for the long-term goodwill, but financially I didn't have any incentive to work on finding a buyer. The chances of us selling for greater than 500k after the team left quickly approached zero. Good startup founders are usually quickly picked up by other companies or launch into another gig of their own. For me, it was a hard decision to pull the plug on EventVue, but once I did, I wanted to move on as quickly as possible. Wrapping up the last pieces of EventVue was painful enough.

The code still lives on my laptop and I reference pieces from time to time. I would have open-sourced it, but it included a lot of expectations about how the infrastructure was set up and would have taken a ton of time to generalize or properly document. The demand obviously wasn't there or else we wouldn't have gone out of business.

So there she lies. ~100,000 lines of code and 3 years of my life. May she RIP.

randy99
I suspect most startup code out there is way too customized for their own purposes/environment. Plus would look like the wild west! (cowboy code everywhere!)
MrMike
I share similar battle scars. As painful as these situations are, the ability to pick oneself up, dust oneself off, and get back into the game is, in my humble opinion, a great skill (as long as you take the knowledge of what worked, what didn't, etc. with you into the next endeavor).
Joakal
I've attempted to make my code as OO possible. This means I can actually reuse it later on other projects. Authentication, ORM, etc.

Everyone should release configurations at least. You can easily release those and those who say you can't either have really bad configuration or too lazy.

joevandyk
FWIW, you can have modular code without OO.
stickfigure
One of the big advantages of deploying on Google App Engine is that it just keeps running by itself, and usually for pretty cheap. Just let it run as a living portfolio.
ghostganz
A few of my old colleagues bought the code from our bankrupt employer and made a second attempt at creating a business with it. They did better than the original owners, but not much.

In all other cases I've come into contact with the code just dies. Open sourcing would be nice, but that's obviously not a priority in a bankrupcy.

mvanveen
Open sourcing is always a good option.

Etherpad (http://en.wikipedia.org/wiki/EtherPad) got sold to Google and the code was subsequently open sourced. It's since launched a number of hosting sites providing collaborative editing services. The source code lives on, there's even a node.js port available with a much smaller footprint (see Etherpad Lite (https://github.com/Pita/etherpad-lite).

Another good example is Mozilla. When Netscape crashed and burned Mozilla somehow managed to come out of the ashes.

jfeldstein2
Make sure you learned from it while you were writing it. Then it never completely goes away.