Kuro5hin.org: technology and culture, from the trenches
create account | help/FAQ | contact | links | search | IRC | site news
[ Everything | Diaries | Technology | Science | Culture | Politics | Media | News | Internet | Op-Ed | Fiction | Meta | MLP ]
We need your support: buy an ad | premium membership

[P]
Open Source Successes and Failures?

By hyssop in Culture
Tue Oct 03, 2000 at 11:50:12 AM EST
Tags: Help! (Ask Kuro5hin) (all tags)
Help! (Ask Kuro5hin)

I'm trying to convince my management to open source a project. They seem to agree that it is a good idea, but they want to be sure they do things right. I'm looking for cases where companies have done a half-assed job of going open source, and have failed. I'm also looking for cases where companies have truly understood the open source community, and as a team they have succeeded.


Here's a list of companies I think were successful in embracing open source: Here's a list of failures:
  • Sun
Here's a list I'd be interested in seeing opinions on:
  • Netscape (now AOL)
  • IBM
  • HP
I'd also be interested in differing points of view as to what companies have and have not been successful (maybe you think Sun was successful, and Redhat wasn't?).

Sponsors

Voxel dot net
o Managed Hosting
o VoxCAST Content Delivery
o Raw Infrastructure

Login

Related Links
o Digital Creations
o Also by hyssop


Display: Sort:
Open Source Successes and Failures? | 40 comments (27 topical, 13 editorial, 0 hidden)
The Failures are Hidden (3.40 / 10) (#1)
by JB on Sat Sep 30, 2000 at 11:05:06 AM EST

A friend of mine did his PhD on failures and successes in implementing Geographic Information Systems. It was almost impossible to get information on the failures - no company wants to broadcast their mistakes, or even admit that they make them. The projects that work are trumpeted in every brochure, web page, and ad-rag. JB

OT: friend's PHD (none / 0) (#36)
by maphew on Tue Oct 03, 2000 at 02:47:36 PM EST

Hi JB,

Can you refer me to the thesis you mentioned? Or pass my addy along to your friend?

Thanks,

matt.wilkie@gov.yk.ca (http://renres.gov.yk.ca/pubs/rrgis/)

[ Parent ]
FuckedCompany.com (3.12 / 8) (#2)
by Jack9 on Sat Sep 30, 2000 at 11:15:11 AM EST

That's where I would look for the failures ^^ A large number of companies OS a project and have expectations. That's how management works. They authorize something and have expectations. OS products which have been successful are those which are general-use tools and proceed, without expectation. The whole "contributed code" and "thousands of eyeballs catching bugs" are all after-the-fact marketing jargon. If you are OSing because of (a) technical, pragmatic, reason you're already in trouble.
Often wrong but never in doubt.
I am Jack9.
Everyone knows me.

Not who but how (4.33 / 9) (#4)
by Skuto on Sat Sep 30, 2000 at 11:30:24 AM EST

First a note:

if you want to be successfull isn't it more logical
to ask HOW to succeed rather than to ask WHO succeeded?

After all there might be external factors involved in
the success of the companys you mentioned. Everybody's
history is different.

I personally like to think of TrollTech as a nice example.
They were able to get their toolkit whidespread within
the open source community and got adopted by a major
free software project. When they realized the licensing
was creating trouble, they gave in (eventually), but
are still able to sell it exactly like they originally
intended. Free software developers are happy because
they have a good toolkit, TrollTech is happy because
of the extra free publicity. What else could you hope
for?

--
GCP

Re: Not who but how (4.50 / 2) (#9)
by hyssop on Sat Sep 30, 2000 at 12:25:31 PM EST

I was going for case studies. I know how things should be run, but convincing my management to put their blind faith in me is another issue. My explaining that I grok the community isn't sufficient. They want some sort of facts. If I can point to a company that tried to go open source, but screwed it up because it didn't understand the community, then I have information that says, "Company A did X, Y, and Z, and succeeded. Company B did not do X, Y, or Z, and failed." They understand history and are prepared to learn from it.

[ Parent ]
Which OS license? (3.88 / 9) (#7)
by FlinkDelDinky on Sat Sep 30, 2000 at 11:58:49 AM EST

Which OS license will you use?

I think it's important. You've got GPL, LGPL, BSD, Artistic, etc. My preference is GPL. A lot of people like the BSD style too.

I think you should stick to a major license if you can, GPL or BSD. Just because you OS something doesn't mean anybody's going to pay attention to your product. There are plenty of great ideas that didn't make it in OS land. I'm not sure why some things flourish and other things die while others linger.

There are plenty of examples. The Linux kernel is wildly succesful, the HURD is not moving very fast, there are all kinds of dead Windows clone OS efforts.

You'd think one of the many Windows kernel efforts would be a major success because everybody wants apps, everbody already knows it and how it should work, etc. Perhaps Wine killed those efforts, or maybe the programmers who actually do the work just aren't inspired by Windows. Who knows?

The HURD is still alive and kicking and I hope to use it someday. Considering the pedigree of the HURD I'm surprised it's not going head to head with Linux already. For the most part the OS programmers love the GPL and from a cultural perspective you'd think the HURD would bleed off talent from other OS kernels (Linux) just because of that. Perhaps it's a 'cult of personality' type thing. With all the commercial Unix vendors making significant code contributions to the Linux kernel it's going to be difficult for HURD to keep up let alone catch up.

Why did Linux succede? I'm not sure. Partly because it was GPL'd? Partly 'cool factor'? Partly because it fed on the minix kernel programming community (attracting them with the GPL and a much more open development process?)? Right place, right people, right time?

Maybe you should define what advantages your software provides its users. Why would those advantages be increased by OSing. Why would a programmer work on this code for free. What license would attract/repel intrested programmers? Who will these programmers be? System administractors, hackers looking for a 'cool' outlet for their programming creativity, technical types within a specific industry (former customers), or something else?

I'm intrested in your answers to the above questions.

Re: Which OS license? (4.00 / 1) (#12)
by hyssop on Sat Sep 30, 2000 at 12:55:29 PM EST

This would actually be several small projects over time. It is my hope that over time, the available source code would spawn other independent small projects which could be started and completed quickly, using the existing source code as a base. I'm not sure how well this will work, as I'm not sure I've seen anyone try this on such a small scale before. But just trying to articulate the problem to you folks is giving me some great ideas.

Given the nature of these projects, I realize that it will be very important to have the proper license. I'm a fan of all the four licenses you mentioned. Since these will be several small projects, it may be that we would use each of them, depending on what sort of life we hope the code takes on.

The programmers that would most be attracted to these projects would be dedicated hackers who would find this stuff cool (GPL, LGPL).

But in some cases, we might want to also attract commercial developers, and allow them to use our code bases and modify them (or use them in LGPL case) as they see fit, for commercial products (LGPL, BSD, Artistic). However, I think that will be a small fraction of the effort.

I'm aprehensive about going into too much detail about the projects right now. I don't want to jinx it, it's still in the "what do you think?" stages. But also, I think it's a great idea, and I don't want our competition to get there first.

Thanks for the great questions, and let me know if you've got more, or if I missed any.



[ Parent ]
Re: Which OS license? (3.00 / 1) (#17)
by Michael Leuchtenburg on Sat Sep 30, 2000 at 06:55:10 PM EST

Well, if you want to be able to sell the changes people not in your company make, use an MPL-style license. You'll lose external developers on that, though.

[ #k5: dyfrgi ]
[ TINK5C ]
[ Parent ]
Aureal Failed (3.70 / 10) (#11)
by iCEBaLM on Sat Sep 30, 2000 at 12:54:08 PM EST

I'd like to add another company to the list of failures: Aureal, former makers of soundcards and now bought out by Creative.

They released only a half open/half binary driver with no specs, and because of that and the fact that they're gone we can't do much to improve it. The project still goes on but is reduced to fixing bugs in the Open Sound implimentation and keeping compatible with newer kernels.

-- iCEBaLM

Re: Aureal Failed (4.00 / 1) (#30)
by Dakkon on Mon Oct 02, 2000 at 02:53:48 PM EST

Yes, but Creative DID completely open-source their drivers for the EMU10K APU and since they now own Aureal's IP, it would follow suit that they would completely open up the drivers/specs for Aureals APU.

Besides which, the fact that Aureal didn't "get" the open-source movement probably had very little to do with them going bankrupt. Much as some might like to think otherwise, Linux and the open-source movement still don't have much market influence, certainly not enough to cause a company to go bankrupt because they didn't open their drivers/specs at any rate.

[ Parent ]
Corel Bad, Progeny Good (3.50 / 8) (#14)
by pretzelgod on Sat Sep 30, 2000 at 03:08:52 PM EST

As a Debian user and (soon to be official) developer, i'd say Corel is an example of how not to do things. As an employee of Progeny Linux, i'd like to think we're doing things right.

Progeny Debian isn't a distribution based on Debian; it is Debian. We're building on the tools that are already there to make installation and configuration easier, and we're providing a stable version of Debian's unstable distribution. Everything we do is freely available and is being offered to the Debian developers (which most of us are anyway) for inclusion into Debian proper.


-- 
Ever heard of the School of the Americas?


what are you actually interested in? (3.40 / 5) (#20)
by sergent on Sun Oct 01, 2000 at 03:42:41 AM EST

Defining "successful in embracing open source" is hard, let's go web surfing.

If you are interested in whether the companies have had business successes with free or free-ish software, why are you asking Kuro5hin? Presumably Kuro5hin is primarily a technical rather than business community and the responses will be not particularly reliable.

So I guess you are actually interested in which companies have the most positive community reputations. It's not clear that this directly translates to whether those companies have built successful businesses--take Cyclic for an example.

Someone smart said something like "it's all about figuring out which 90% of the community you can afford to have hate you" (paraphrased; not sure of the original source).

I mean, personally, my impressions of the companies you list are rather different... I'm not much of an ORA or Sendmail fan, for example, and I had only heard of Digital Creations once or twice.

So I would posit that all of these companies have annoyed some segments and delighted other segments, and your list just locates you in a particular segment. Your company needs to figure out who they want to make happy, and if they want to emulate someone, they hafta make sure it's someone with the same target audience.

More meat, and another kind of 'failure' (3.62 / 8) (#21)
by mr on Sun Oct 01, 2000 at 08:52:13 AM EST

1) The article can use more 'meat' to it.

2) There is a failure mode not mentioned here.....the BSD OS. Many have taken code form it (good enough to take code from), many run very big sites on it (cdrom.com, yahoo, etc), and yet it can't seem to get out of the shadow of linux.

Are you looking for technical excellence via open source, or marketing success? And, are you prepared to have your code bundled with Linux (or BSD) and have some one refer to your code *AS* xxx Linux or xxxBSD because they bought xxx linux/BSD and your code was there.

Ask Jordan Hubbard, and he'd say that BSD is *NOT* a failure. Ask Linus T, and he'll have a differnt view in public vs private.

Re: More meat, and another kind of 'failure' (3.50 / 2) (#29)
by bsletten on Mon Oct 02, 2000 at 12:04:16 PM EST

Ask any of the several million BSD users and they'll probably chafe at the term "failure" too. Just because Linux took off more quickly does not mean other alternative OSes are failures. An adoption slope > 1 cannot be termed a failure.

[ Parent ]
VA Linux (3.50 / 4) (#24)
by epsilon on Sun Oct 01, 2000 at 10:34:00 PM EST

VA Linux is a pretty successful company from what I hear. You should add that to your list.

"Success" and "failure" in ope (4.40 / 5) (#26)
by Aquarius on Mon Oct 02, 2000 at 05:08:56 AM EST

I'd be interested in how people see "success" and "failure" in open-sourcing being evaluated. Saying "we've succeeded in embracing open source" could mean any one or more of:
  1. We use open source software
  2. We have open-sourced some software
  3. We have open-sourced some software and the software is now better due to that open-sourcing
  4. We have open-sourced some software and our revenue stream has consequently gone up
    I think that quite a few companies would count (1), which is obviously a somewhat minimal example. (2) is a move closer to what I would say that "embracing open source" means, but it was conceivably an ill-advised step, given that (3) or (4) are what should be happening from a business perspective.
    The ideal thing, from a business viewpoint, is (4), since businesses are in business to make money. However, (3) is indirectly going to get you (4), the business might well hope (all cynicism about marketing over content aside), so it's also a laudable goal. (2) might mean that the company jumped on the open-source bandwagon without contemplating their move in depth; just going open-source for the sake of it (i.e., without a business case including tangible estimated business benefits (whether financial or measured in more ephemeral things such as market goodwill)) is bad for a business, not good.

    Aq.


    "The grand plan that is Aquarius proceeds apace" -- Ronin, Frank Miller
Re: "Success" and "failure" in (4.00 / 2) (#35)
by your_desired_username on Tue Oct 03, 2000 at 02:37:51 PM EST

The most likely meaning of "we've succeeded in embracing open source" is actually

5. We are trying to appeal to an open-source-liking market segment.

[ Parent ]
Cynical but true (none / 0) (#39)
by Aquarius on Wed Oct 04, 2000 at 04:27:13 AM EST

In my initial comments, I included the "we're doing this just to look good to the open-source crowd" as part of "we're using open source products", but you're right to split them in two.

The issue here is possibly that Open Source as a concept is a bandwagon. Some companies may dismiss it, some may "embrace" it (with the definition of what exactly that means dubious, as we've already discussed), but most IT personnel are at least peripherally aware of it. The flaw here, as I see it, is that companies are jumping on the Open Source bandwagon without buying into the "everyone can be a developer" concept. If five people use a piece of software, and all of them are prepared to report bugs[1], then you've got a good user pool. If five thousand people use a piece of software, and the same five only report the bugs, then you've potentially got 4995 disppointed users who complain that the software is buggy, which makes it just the same as other non-OS buggy stuff. If OS is being pushed as robust (which it is; stability and features seem to be seen as more important to IT than cost, almost, since most IT departments are resigned to paying Big Money for the software that they need), then it needs bug reports from users, because that's how the OS model works. ESR talks in The Cathedral and the Bazaar and the followup papers about leveraging your users to help development in some way. He reiterates time and time again that this way leads to better software, debugging is parallelizable, all bugs are shallow to somebody, and so on. However, if all the users don't buy into that and instead expect to get robust and beautiful software, for free[2], without aiding in the development process, then the model falls apart.

Anyway, the point of this ramble is that companies who jump on the OS bandwagon without buying into the concept may well turn around in a few years, when the sheen of "embracing Open Source" has worn off, and say "hey, all this software is really buggy and still under development", and then depart. That's why the somewhat cynical (but true) view above that "embracing Open Source" can mean "we're cool because we have one FreeBSD server, look at us, buy our product" is a bad thing.

Aq.

[1] I don't consider that users of OS should necessarily consider actually fixing code problems "buying into the concept". I don't fix code problems, because C is not what I do. I'll happily report bugs, though.
[2] beer


"The grand plan that is Aquarius proceeds apace" -- Ronin, Frank Miller
[ Parent ]
What about Cygnus? (3.00 / 3) (#27)
by sbeitzel on Mon Oct 02, 2000 at 09:39:13 AM EST

I don't know how they're doing nowadays, but certainly they made enough money by supporting OS software to pay full-time staff in Scotts Valley (Santa Cruz? Can't remember.) for several years. I still install cygwin on every Windows machine I have to use.

Re: What about Cygnus? (none / 0) (#28)
by afc on Mon Oct 02, 2000 at 10:21:16 AM EST

Nowadays, as most everybody who hasn't been living in a cave knows, they're a wholly owned subsidiary of Redhat Inc.
--

Information wants to be beer, or something.
[ Parent ]

DOH (none / 0) (#40)
by sbeitzel on Wed Oct 04, 2000 at 10:04:24 AM EST

OOG TOLD LEM SOMETHING ABOUT THAT, BUT LEM NOT REMEMBERING. LEM STILL FUZZY-HEADED FROM OOG'S CAVE WEED.

Yeah, I live in a cave and don't get out much. The big room with the yellow eye is a SCARY place. What's your point?

[ Parent ]
What about SGI? (3.50 / 4) (#31)
by lovelace on Mon Oct 02, 2000 at 03:31:34 PM EST

I think SGI is a company that has done pretty well open sourcing various things (XFS, OpenGL SI, Open Inventor). I interned at SGI this past summer and worked with the group responsible for the open sourcing of Open Inventor. Rather than hold onto a project that wasn't doing anything at all (no engineers were assigned to it), SGI released it under the LGPL so that people could do what they wanted with it. If every company did that, we could have some pretty cool stuff out there. What if Corel decided to Open Source Word Perfect? (Can anyone think of others?)

Don't think, however, that they only open source dead projects. SGI has also made available under the GPL a set of 64 bit optimizing compilers (i.e. for the Itanium). They've also done a bunch of other stuff (which you can see here. They may not get as much press as other companies, but I think SGI is definitely a company that understands open source and because of that has succeeded.

Not much to lose (3.33 / 3) (#32)
by stuartf on Tue Oct 03, 2000 at 12:40:40 PM EST

Seems to me that the high profile companies that have released open source products are the companies that have nothing to lose e.g. Netscape (browser failing miserably, losing market share). The only large scale successful open source projects I can see are Linux and the Internet products (BIND etc). These products are differentiated (IMO) by the fact that they were never intended to make money.

How not to do it: Mozilla (3.50 / 2) (#33)
by noahm on Tue Oct 03, 2000 at 01:06:55 PM EST

I want to cite Mozilla as an example of how not to open a project. I don't claim that Mozilla as a project is a failure (in fact, I'm using it right now and it's getting extremely good). My problem is with how Netscape went about opening it. The initial code release of Mozilla was severely broken. It was not the code for the most recent version of Navigator; it was a gutted shell that barely compiled. I realize that Netscape had to do much of the gutting to remove code that they'd licensed under non-disclosure terms, but when the first adopters of an open source project have to spend all their initial development time fixing a broken code base, then the project is in for an uphill battle.

In Mozilla's case, the recovery was impressive, but still slow. Many other projects would not have succeeded at all.

noah

Re: How not to do it: Mozilla (3.00 / 1) (#37)
by Matrix on Tue Oct 03, 2000 at 03:57:50 PM EST

Actually, IIRC, Mozilla eventually wound up completely throwing away the Netscape code they were given and starting from scratch. But I will admit that Netscape did a horrible job of opening their software (I'd like to think it was because of AOL's influence). And the license they chose to use didn't help much at all.


Matrix
"...Pulling together is the aim of despotism and tyranny. Free men pull in all kinds of directions. It's the only way to make progress."
- Lord Vetinari, pg 312 of the Truth, a Discworld novel by Terry Pratchett
[ Parent ]

The big question: Why? (3.00 / 3) (#34)
by mandomania on Tue Oct 03, 2000 at 01:10:05 PM EST

This is the $25,000 question. Why is your company looking to open source this project? If it's just a PR ploy to draw some free publicity, then don't bother reading the rest of this comment. Just slouch back into your faux leather chair and sip your overly-priced-yet-still-crappy coffee. If you want/expect your software to change the world, then read on fellow starry-eyed maniac :)

Assuming your project is high-quality and useful (2 REALLY big if's), it will be used regardless of it's openness. However, if you want your project to grow in utility and quality, open it up. If it's used, it'll be tinkered with. All you have to do is look at some of the big names in the open source movement (Apache, Sendmail, BIND, etc.) to see this. The key here, however, is that the code was sound to begin with. No one's gonna take crappy code and fix it. They will, however, take good code and make it great.

--
Mando
The Code is Sound.
OSS politics? Why so much evasion of the topic? (none / 0) (#38)
by eventi on Tue Oct 03, 2000 at 05:01:30 PM EST

It seems like people are curiously evasive of the actual topic of the article. And here I go, ignoring it myself:
The direction of this discussion, or lack thereof is analogous to the OSS movement itself. Too many chefs, and not enough cooks.

Open Source Successes and Failures? | 40 comments (27 topical, 13 editorial, 0 hidden)
Display: Sort:

kuro5hin.org

[XML]
All trademarks and copyrights on this page are owned by their respective companies. The Rest 2000 - Present Kuro5hin.org Inc.
See our legalese page for copyright policies. Please also read our Privacy Policy.
Kuro5hin.org is powered by Free Software, including Apache, Perl, and Linux, The Scoop Engine that runs this site is freely available, under the terms of the GPL.
Need some help? Email help@kuro5hin.org.
My heart's the long stairs.

Powered by Scoop create account | help/FAQ | mission | links | search | IRC | YOU choose the stories!