Hacker News Reader

Comment by baggachipz | Hacker News Reader

I might be feeding here, but the lesson is that you can't count on things Google releases to be around very long. Even APIs for popular products, like Google Maps. Just don't expect anything they release to have a long shelf life. It's annoying and it sucks, and it is now laughable when they tag things with phrases like "free code hosting forever" and so on.

For them to change, they'd have to either: 1) Support everything into eternity, which given their propensity to release new things frequently would grow to be impossibly expensive, or 2) Be more disciplined in what they release, so that it can be supported in perpetuity. I don't see either happening, so we end up where we are now. They have to keep innovating or they'll stagnate and die.

View on HN

They kind of get a bad rap for this because of how large and visible they are. The funny thing is that many of the people here on HN have worked on projects that had very finite lifetimes as well, often closing them down for some of the same reasons.

Someone has to invest the time and money to keep services running, even if only limping along. Google likes to experiment and test the waters. It's part of why they've been so successful. I hope they keep killing off their underperforming/underused services so they can keep on pushing other things forward.

However, closing down stagnant projects is very different from locking down access to previously open APIs (Google Maps). Since we're discussing the closure of Google Code, I won't get distracted with that controversial topic.

I also find a certain level of humor in the oft-espoused "release and pivot" ideology, accompanied with the pleas for mercy found in these Google product shutdown threads. What is different? Why is Google not supposed to release and pivot like everyone else?

It is puzzling, but I think their rapid ascension has led to some people setting their expectations for Google too high. Others seem to hold them to some golden moral standard (of their own design), for some reason.

In the end, they are a for-profit business. Not a charity.

At least from my perspective, the problem is that Google sometimes releases into already-existing fields at a price point no one can compete with, absorbs the entire market, then pivots away. This is pretty destructive behavior, and I think fundamentally different from entering a field in which you intend to try to run a business (as opposed to subsidizing your entry into the field with profits from elsewhere in the company in order to undercut current players).

Except in this case they didn't absorb the market and then pivot away. The market moved away from them and consequently they shut down a service that nobody really uses anymore.

Sorry, should have made it more clear that I wasn't talking about Google Code specifically, but was responding to the parent comment regarding "release and pivot".

I think it is because of the side effect (can it be called an externalized cost?) to developers wanting to use existing open source code and not being able to find it. Google code is used by a lot of people.

Whereas an early start-up with 100 users selling a product with not much lock-in can pivot without too much side effect. The users know they are dealing with a start-up and may expect it.

Because at some point you stop being a feature of the landscape, and you become the landscape - or a big part of it, at least.

That's where GitHub is now. If GitHub closed in the same way, that would be... unfortunate.

That seems to be where Google Code was trying to go. For whatever reason, it never quite made it.

But if you're part of the scenery - and Google Code was, for a while - and not just a shack hardly anyone visits, a burden of responsibility goes with that.

Yes, and the projects that people on HN have worked on that have had very finite lifetimes get a very negative reaction on here too, for much the same reason. In fact I regularly see people suggesting that it reflects badly on startups as a whole and that you shouldn't trust them with your data whenever one pulls a stunt like this.

> Support everything into eternity

'Eternity' overstates it. Plenty of vendors support projects long after they are obsolete. IBM supported OS/2, released 1992 and a marketplace flop, until the end of 2006.

That means that if I need something to be around for awhile, I'll count on IBM and not Google. Google has the perogative to make that choice, of course, and long-term support generally is associated more with the business market than with consumers. Party that's because it costs more for businesses to change, partly IMO because consumers are poorly educated customers.

EDIT: It's a little odd to see this very normal comment modded down to -1 (now -3!). Another perfectly fine comment was modded down to 0. I don't care about the points at all, but it supports my instinct that some on HN are trying to protect Google. Employees? Fans?

> IBM supported OS/2, released 1992 and a marketplace flop, until the end of 2006.

That's because users paid for it.

Google gave the world high-quality, free open source code hosting for eight years. I don't see how that's anything but a net positive.

I think there's a huge difference between continuing support of a product and continuing support of a service. Not the least of which is that products can pick up their own life and unofficial support even after the company has abandoned them. The best you could hope to do for a service is faithfully duplicate it, with all the effort that involves.

And for OS/2, IBM even allowed it to live on by licensing it to a third party:

https://en.wikipedia.org/wiki/EComStation

> 'Eternity' overstates it. Plenty of vendors support projects long after they are obsolete. IBM supported OS/2, released 1992 and a marketplace flop, until the end of 2006.

I'll give you three guesses as to how much those companies still running OS/2 were paying for support in 2006.

Have you ever done business with IBM? They do the exact same thing, all of the time, except that it's software so you need to deal with migration.

When IBM buys a company/product, they will merrily throw everything out and declare that v.next is the next version of whatever product you are using. Then the legacy product goes in life support.

OS/2 lived on because it was widely deployed in banking for ATMs and POS, and it generated a lot of complimentary revenue. So when a company like Target had 100,000 cash registers, IBM got to sell services around deployment and maintenance, and management software like Tivoli to manage the devices.

you can't count on things Google releases to be around very long

code.google.com was around for 7 years. What more do you want?

FWIW, it was 8+ years :)

It was launched at OSCON in 2006.

Why limited to google? You can't count on anything being around forever. But that is hardly revelatory.

I think people tend to hold large companies with far-reaching services to a higher standard in this regard. I'm not arguing that stance's validity, but if MS or Google or even Facebook (yes, I know) release a product or API, people assume that it will be well supported. Partly because of $large_company's deep pockets, and partially because it's assumed that those companies understand the far-reaching consequences of their actions and want to act responsibly.

> partially because it's assumed that those companies understand the far-reaching consequences of their actions

The large companies are made of small clueless people just like every other company on the face of the Earth.

There is two alternatives that I can think of.

(a) For things (eg. Reader, Google Wallet for Digital Goods) that could be viable businesses in their own right with some modifications, spin them off as small companies.

(b) Open source the project they no longer have interest in. (as was done for Google Wave, which arguably wasn't necessary because the level of interest in Wave at that point was quite low)

In this case (and perhaps they already have this), I would suggest that they allow anyone to discover all the projects on Google Code and export them such that they could be imported into other code hosting sites, allowing everyone else to build a Google Code alternative.

This is the problem with proprietary cloud services. As long as they're hosted by and run by a single company, you can't be sure they'll stick around.

Yeah, like http://www.codehaus.org/ is living forever? ;-)

Reply on news.ycombinator.com