Red Hat to Coax Code Contributions From Companies

Stony Stevenson writes "New Red Hat CEO Jim Whitehurst has hit out at enterprises, bemoaning that billions of dollars are wasted each year because 95% of companies won't share code. Speaking at the Open Source Business Conference in San Francisco, he said his company must take a larger role in urging enterprises to participate in open source projects and, in some cases, coax code contributions out of companies that have made in-house improvements. He now feels Red Hat should lead the way 'It should be part of Red Hat's job to define development in a new way, and get companies to work together' on shared projects, he said. The joint development projects would be designed to cover non-competitive parts of an industry, with individual companies still focused on their own competitive business applications."

Comments

Cable code?

[robipilot]

My first read of the title was WHAT? Code for coaxial cable? Me no get it.

Re:Cable code?

[Red Flayer]

Yeah, well, I'm urging all companies to cat-5 their code, that's gotta be better then coaxing it.

Yes, but...

[MrNaz]

While I agree with Jim's sentiments being an Open Source advocate and all, I think Red Hat has no right to attempt to coax or coerce companies into giving away code. If OSS is the future, then it will happen, with or without Jim's little tantrum.

It is ridiculous for a CEO to attempt to paint his company as some kind of inspired model upon which other companies should remodel themselves. Aside from being futile, attempting to turn the Old Establishment around does nothing but hurt the nascent organisations that will make up the New Establishment by casting doubt on their methods and making them look like they are non-viable without the support of the Old Establishment. I can see Ballamer right now, in a room full of beaureaucrats saying "See? OSS is all about getting handouts to survive." Furthermore, it is brining wolves in amongst the lambs.

If Jim wants to make a difference, he should fund new development from emerging pools, like Google with the GSoC (not that I'm a Google fan, but that's another story), or IBM with their paid employee time contributions, or EnterpriseDB with their backports to the PostgreSQL team or Sun with their (somewhat clumsy) contributions to the OSS community. There are plenty of companies already doing what he says, he should be happy for that and encourage those already willing rather than attempting to project an agenda onto those it does not suit.

Having a whine that companies in the Old Establishment should be putting free money into his playpen is a naieve, futile and potentially harmful thing for Jim to be doing. It'd be better all round if he put his money where his mouth is rather than asking others to put their money where his mouth is.

Re:Yes, but...

[99BottlesOfBeerInMyF]

I think Red Hat has no right to attempt to coax or coerce companies into giving away code.

This is a straw man argument. The article said "coax." The summary said "coax." You added "coerce" which is not something anyone had brought up. In principal it is no different from saying that Redhat has no right to attempt to coax companies into giving away code or molest children.

If OSS is the future, then it will happen, with or without Jim's little tantrum.

I strongly disagree. Microsoft spends a lot of money convincing purchasers that they are better off buying all Microsoft, proprietary solutions. At the same time, not a lot of people making purchasing decisions understand the OSS business model and how it can save them a lot of money. Providing a voice that explains and advocates this method is very useful.

It is ridiculous for a CEO to attempt to paint his company as some kind of inspired model upon which other companies should remodel themselves.

He's not "painting his company" as a model. He's advocating an alternative development method that differs significantly from classic economic models. Redhat has done well by being a contributor to that model. That is not ridiculous at all.

Aside from being futile, attempting to turn the Old Establishment around does nothing but hurt the nascent organisations that will make up the New Establishment by casting doubt on their methods and making them look like they are non-viable without the support of the Old Establishment.

Old Establishment, New Establishment?!? Redhat is simply talking to companies, whether new or old, and trying to sell them on a cheaper way to do business that also helps undermine software lock-in strategies. OSS is, quite simply a feature of software, that many do not appreciate the advantage of. It needs to be explained, like most other new features consumers are not used to using.

I can see Ballamer[sic] right now, in a room full of beaureaucrats[sic] saying "See? OSS is all about getting handouts to survive." Furthermore, it is brining[sic] wolves in amongst the lambs.

In such a meeting, Ballmer is a salesman, and most companies don't trust salesmen. Microsoft already tries to paint OSS as something that is risky and unusable to big business, but not too many people are believers, given that IBM argues the opposite.

If Jim wants to make a difference, he should fund new development from emerging pools...

There is a lot of software in use today which is used in various niche applications. Quite often such software is custom built for a company, and their competitors also use custom built software. This software is not really a point of competition between these companies, just something they need in order to do business. What Mr. Whitehurst is saying is that Redhat can be more proactive in going to these companies and getting them to open source this code and allow all the companies that need that niche application to share the development costs, rather than each of them paying to develop their own version. This leads to many advantages for the companies including: lower overall development costs, more competitive bidding on development, and standardization within the industry for interoperability. Further, getting some of this code open sourced gives Redhat (and other such companies) a way to undercut proprietary software developers when providing custom coding, support, and added services.

There are plenty of companies already doing what he says, he should be happy for that and encourage those already willing rather than attempting to project an agenda onto those it does not suit.

I think you're still missing the point. This is about evangelizing OSS as a way to cut costs for companies that currently don't understand or contribute to it. There is a huge, potential market for OSS development and a lot of closed

Re:Yes, but...

[gdek]

A comment this ignorant, and yet this highly rated, pretty much demands a rebuttal.

1. "I think Red Hat has no right to attempt to coax or coerce companies into giving away code. If OSS is the future, then it will happen, with or without Jim's little tantrum."

Guess what? There are *a lot* of companies coming to Red Hat, right now, *asking how to participate in open source projects.* So Jim is not talking pie-in-the-sky here; he's talking about capitalizing on momentum that already exists. There's pretty much zero coercion involved here.

2. "It is ridiculous for a CEO to attempt to paint his company as some kind of inspired model upon which other companies should remodel themselves."

So why is it, exactly, that Sun and Novell are trying to rebuild their business models, again? Help me out here.

3. "If Jim wants to make a difference, he should fund new development from emerging pools, like Google with the GSoC (not that I'm a Google fan, but that's another story), or IBM with their paid employee time contributions, or EnterpriseDB with their backports to the PostgreSQL team or Sun with their (somewhat clumsy) contributions to the OSS community. There are plenty of companies already doing what he says, he should be happy for that and encourage those already willing rather than attempting to project an agenda onto those it does not suit."

Considering that *every engineer at Red Hat is an open source software engineer*, either full-time or part-time, I'd say that Red Hat is funding plenty of open source development all around, thanks very much. Or maybe you don't think that any of this stuff counts.

4. "Having a whine that companies in the Old Establishment should be putting free money into his playpen is a naieve, futile and potentially harmful thing for Jim to be doing."

As it turns out, executives at big companies are smarter than you are. See, they understand the difference between "differentiating value" and "non-differentiating value". (Read some Bruce Perens if you don't get that idea.) Jim Whitehurst was the COO of a Very Large Company that had a larger annual IT budget than Red Hat's entire annual revenues. He saw firsthand how much money and manhours IT departments waste on software that doesn't actually add any value to the business. "Old Establishment" is looking desperately to make sure that those IT guys are building value, not wasting time on stuff that doesn't differentiate them from their competition. Understanding *and participating in* the open source model is one of the best possible ways to do exactly that. Which is why "Old Establishment" is coming to Red Hat and saying "help us".

The limiting factor is that Red Hat is not yet big enough to provide all of the services and guidance that these customers need. Jim is committing himself, publicly, to meeting that challenge. At Red Hat, we're all very proud of him for saying so.

Re:Yes, but...

[radagenais]

I think its fair for RH to position themselves as a leader in the industry and, at risk of getting flamed, I humbly submit that, overall, they have contributed positively to OSS.

But I think that Jim's aim might be a little off. He points to enterprise, but I think that there is a massive swath of small to medium sized solution providers who are hording their code when they build enhancements for customers. This is their little cachet, their angle on the (primarily local) market, their "solution". A number of times I've pointed out to consulting firms I've worked for/with that they weren't compliant with GPL because they weren't putting their code improvements back into the wild, and they looked at me blank-faced, "Isn't it free??" "Sorry, boss, that's the BSD license. This is GPL. You gotta share." A frequent example that comes immediately to mind from a couple years back is Asterisk solution providers.

As for enterprise, you need to show them value in the form of professional services. If they can get expertise and help, they will be open to play ball. This is an area that RH can show their strength as a services company. If Jim puts his money where his mouth is, it could work.

My own focus is on professional services and I perceive OSS as a great opportunity to 1) improve the quality (security, interoperability, all that) through sharing of knowledge, which is just good science; 2) improve the professional services opportunity for Slashdot types. Services should be the biggest piece of the pie, not hardware or licensing, and this will help elevate the profession as well.

As a post-script, I think that developers have a certain amount of professional responsibility to point out the licensing model for any code they seek to build into their solution. If this was discussed more often, it might enhance awareness, dissolving some of the misconceptions about OSS. People aren't going to decide to share their code in hind-sight. You want to get them involved from the start of their project. I think Jim gets this and he's got to speak to it now and form customer partnerships to get that rolling.

There is no doubt that this is also a business development tactic for RH, but I see nothing wrong with that.

Re:Yes, but...

[99BottlesOfBeerInMyF]

Where's the advantage to the company that does the initial software development?

I think I already covered that, but here it goes again. Money spent is spent. You can't un-spend it an no one who went business school should fall prey to the fallacy of throwing good money after bad. In general, all companies have already invested in some niche software. The company open sourcing code may or may not have the best software out there, but making it OSS provides them, the users, with a new feature.

When you open source some project you benefit in numerous ways. First, you get are likely to get free code contributed by others (money you might otherwise have had to spend). New market entrants are more likely to adopt the same software you're using, giving you an advantage in interoperability in the future. It is not uncommon for the entire industry to slowly move to the software you already use which is a migration cost for them and if someone else beat you to supplying their code as the open source version you may well find yourself paying money in future to interoperate with them and everyone else or to migrate to their software (which is the new industry standard). Depending on how you release it, you may well have more influence in the development process than others and your competitors incur expense if they try to fork your project to better suit any of their needs which you don't share. If your software does gain market share because of open sourcing it, more development houses will become familiar with it, which means for the next contract to add to it, you can take competitive bids, slashing your costs.

And I'm sorry, but every piece of software is a point of competition. If one company can save money by using something as simple as a better email client, that's a competitive advantage over other companies that don't use the better email client.

Bullcrap. There is tons of software out there, like e-mail clients, which have little or no impact on how competitive a company is. There are just too many available that all meet your needs. The point of OSS development is to share code for areas that are outside your core competency. For example, I worked at a company that made core network routers. Routing software was our core competency, but OS development was not. By using NetBSD or Linux as the basis for routers, we save a lot of money and development work. We contributed to NetBSD and Linux when we needed a new feature or there was a bug that needed fixing. Did our competitors benefit from that? Yes. Did we benefit when our competitors did the same? Yes. Did companies that relied on expensive, proprietary OS's they paid for or made themselves crush us? Hell no. They went out of business or switched to the same OS's we were using because otherwise we undercut their prices by a huge margin. We never open sourced our core routing protocols and code (where we really competed) and the rest we used OSS (OS, dev tools, apache for Web interface, etc.). The same situation applies in almost any industry.

Neither you, nor anybody else in this thread has expressed a single, logical, positive financial reason why a company should open source software that they develop in house.

Yes, I did. You do it to cut future development costs and interoperability costs. It works. It has been working in many industries for decades. I'm sorry if you can't wrap your head around why the business model is a benefit, but I can't be much plainer. There are a lot of companies using Linux, Apache, MySQL, GCC, and hundreds of other projects and making more money than they would otherwise be able to. They also pay money to contribute to those projects as they need to, and it is still a lot less than what they would be paying to maintain an internal, proprietary project to do the same thing.

Redhat absolutely can show a lot of companies the numbers on this and examples from other industries. It is a chance for Redhat to grow their development and support services business, while at the same time costing less than proprietary companies or internal projects. OSS is smart business and first mover advantage has real benefits.

It's worse than that.

[MichaelCrawford]

I did a consulting gig a while back, whose contract specifically said I was not to include open source code in my work for them.

There was no mention of licenses; open source licenses include the MIT and BSD licenses, and many similar licenses that permit keeping the source to derivative works closed. And in fact, Microsoft itself uses a lot of BSD code in Windows, without sharing any of its source.

I was very unhappy about signing such a contract, but I needed the work.

I never really asked why they wouldn't even allow source under the MIT or BSD licenses. I expect that it was a lack of education. If that's the case, I expect their attitude is not uncommon, and sorely needs to be corrected.

For what it's worth, my current employer (I'm no longer consulting) releases the source code to its Linux and BSD drivers as open source, with their source code being provided on our installation CDs.

Yes, but you're appealing against

[drunkenoafoffofb3ta]

how businesses usually think. Share their stuff with others? Give other companies an advantage that WE paid for? NEVER! So yes, it's a huge waste. But you'll have a hell of a time convincing them to change. Um, imho.

Re:Lead the way

[Anonymous Coward]

http://www.press.redhat.com/2008/03/19/source-code-for-red-hat-certificate-system-released/

No Thank You

[Hyppy]

After seeing the absolute filth that is spewed out of most corporations' in-house "development" teams, I'd be very wary of this.

Re:No Thank You

[AshtangiMan]

I agree. The difference that I see is code written within a corp not as a part of the OSS movement is developed with deployment in mind, not with the attitude that others are going to also use this code. This leads to poor documentation (esp in code commenting) and generally sloppy coding. Now, OSS may not be better, but I would hazard a guess to say that it is. Writing code that you know other coders are going to use in other applications/ projects as a matter of pride would lead to better organization, commenting, etc. Of course I may be completely wrong . . .

Re:No Thank You

[kellyb9]

I think this generally has to do with the function of the code. For example, most corporate "in-house" solutions are pressed for time and resources. Many of them are poorly documented as it is often treated as a luxury to have good documentation. Because of the nature of OSS, it typically has better documentation. Most people designing OSS hope to pass if off and allow others to build onto what they've done.

coax yes coerce no

[davidwr]

Perhaps a better word would be to encourage or evangelize. Coercion should have no place in business and the word coax can mean either to benignly encourage or to coerce.

Re:coax yes coerce no

[MrNaz]

You cannot steal what I am willing to give to you for free.

Re:coax yes coerce no

[zotz]

"My point stands even if Jim had said "pretty please with a cherry on top" while wearing a pink hoola skirt."

I don't think it does when it comes to coaxing. I am not sure it does in any case.

If my approach to coaxing someone is to point out to them how they will benefit by doing what I suggest and then they decide to do it... You have a problem with that?

I am interested in Free Music as well as Free Software. When people are afraid to try it with their own music, I suggest they at least experiment. Release a single song that they think of as having good quality under a Free license. Promote it. See what happens. Unless they think there is a good chance they are going to be a one hit wonder, and there is a good chance that they will pick their only potential hit to put under a Free license,there is little risk in such an experiment.

Or I suggest that they start with someone else's Free and copyleft lyric or tune and build on that. This lowers their risk even more.

Baby steps if that is what it takes.

I think that the people to encourage are industry associations. Let them find a way to support Free Software to the benefit of their members.

all the best,

drew
http://packet-in.org/wiki/index.php?title=Main_Page
Packet In - net band, libre music, sometimes gratis.

Re:coax yes coerce no

[Hatta]

You can't rape the willing.

Re:Job loss

[langelgjm]

If companies start sharing code, there will be less code that needs to be written in-house, which means some people are going to be losing their jobs. I'm sure they'll be really thankful to Red Hat.

I already moderated in this article, but I'm willing to lose the moderations just to reply to this.

Analogy: if universities start sharing research, there will be less research that needs to be done in-house.

Um, yeah. Unnecessary duplication of effort is wasteful. Yeah, they could lay off people, or you know, they could use the same number of coders and now accomplish more tasks.

no Job loss

[mcrbids]

If companies start sharing code, there will be less code that needs to be written in-house, which means some people are going to be losing their jobs.

That is "fixed pie" thinking. Underneath your statement is an assumption: that there's only a fixed amount of work to be done, that the amount of work "pie" available is fixed and unchanging. That simply isn't true.

The real purpose of a job is to generate wealth. Janitors create the wealth of a cleaner environment. CEOs create the wealth of a smoothly running organization. Factory works create the wealth of manufactured goods. And so on...

If wealth gets generated more efficiently, everybody benefits, because there's more total wealth to be distributed. An organization that "eliminates" a few positions is then wealthier, which then makes it more likely to increase its product base, thereby creating more positions. While there are cyclical deviations and occasional abuses, (generally covered by existing laws) it's largely a self-regulating system.

Don't be afraid of change. Be afraid of stagnance.

Re:Job loss

[idlehanz]

Or if you're the person cutting the checks, you give yourself a bigger bonus and call it a day (optimistic cynic). Optimistically the company will create a new product and assign the idle workers to this task and generate more revenue. Singing of kumbaya and hugging to follow. Or, the company lays off the extra people UNTIL they create a new competitive product, then hire people to support the new product. Greater disruption, but hey, that's the marketplace. As a bonus we get more time on the X-Box while we wait for the market to correct.

Transifex

[Nushio]

Disclaimer: I am currently a Fedora Translator

Fedora currently uses Transifex, which makes all translations go Upstream, thus sharing what we've translated, with other Software Projects.

I'd sooner share their herpes

[giafly]

... than the code produced by most teams.

Re-use is not just about shoving code on a server and letting people copy it. You also need design, documentation, comments, testing, and ideally some level of support.
A lot of in-house code comes with none of these and as a result is worthless.

Re:Job loss

[s20451]

Analogy: if universities start sharing research, there will be less research that needs to be done in-house.

Your analogy is flawed, because universities do not consume the research that they produce, and they are (usually) not expected to make a profit.

Also, it says right in the summary that "billions of dollars" are wasted on duplication. One obvious way to save that waste is to fire programmers and freeload off of the code of others. I can't think of a good reason to believe that the distribution of the savings will be equitable.

Competitive Advantage

[jamesl]

If I'm the CEO of a big-ass Insurance Company, Bank, Airline or Widget Manufacturer and I just invested a bajillion hours of developer time into creating software that gives me an advantage over my competition, why would it be in my best interest to give my code away?

Re:Competitive Advantage

[Nushio]

If you're the CEO of a big-ass Insurance Company, Bank, Airline or Widget Manufacturer and you just invested a bajillion hours of developer time into modifying (Free?) Open Source Software, it would be greatly appreciated if you contributed back to said projects.

Re:Lead the way

[tux_deamon]

That's because it uses an Oracle backend. They're working on replacing that component, and soon Sat will be opened. http://arstechnica.com/news.ars/post/20070130-8737.html

Bullshit

[DogDude]

The joint development projects would be designed to cover non-competitive parts of an industry, with individual companies still focused on their own competitive business applications.

No such thing as non-competitive parts of an industry. If two companies say, make toilet paper, and one of them has a custom program that let's say, saves energy by turning off unused lights in their buildings. That company saves money on their power bill. That is still a competitive advantage over the other company, even though it has nothing to do with the industry. Why would the company that developed that give that to a competitor, and allow that competitor to improve their bottom line? Every piece of doing business is a competitive advantage. There are no insignificant parts of any business.

Re:Bullshit

[Tranzistors]

Not everyone is a competitor. There may be some companies that are your allies. If you have energy saving software and your distributor uses it, this makes your paper cheaper. It also makes other company paper cheaper as well, however, more costumers now can afford it (or afford it more).

Also, if over all economy improves, the chances are your paper business will improve. Your supplyers can deliver more cheaply, your clients can pay more.

And to finish, if you release reasonably good and useful code, other altruistic companies may contribute back, thus reducing time and effort.

Re:Job loss

[Sebastian Reichelt]

Or we'll find new problems to solve instead of reinventing the wheel independently in a thousand silos.

If only software development was like that! If a software solution for a given problem exists (closed or open source), that still does not mean the problem is solved once and for all. The implementation is hardly ever isolated from the rest of the "product". Even in the open-source world, we still have these large, monolithic programs, and nobody has a chance of reusing just a certain aspect of them. In fact, in some cases it's even worse than with commercial software: Just think of all the applications that exist twice: Once in a KDE version and once in GNOME version.

Until all code is truly reusable and free of everything non-problem-related, programmers will reinvent the wheel over and over again.

Re:I wouldn't say they're "wasted"

[Hatta]

This is the broken window fallacy in action. You might as well suggest that companies hire men to dig ditches all day and fill them back up, just so they can get a paycheck. Rewriting the same code all the time is just as pointless.

If these companies didn't need to waste (yes waste) that money on that code, they could spend that money in other ways. Maybe it wouldn't get spent on code, and there would be less of a market for programmers. But there would be a greater demand for other services, so the economy as a whole would be ahead.

It's not a one way process

[Nicolas MONNET]

If you contribute back to a F/OSS project, such project grows and attracts new contributors, who will in turn give you stuff for free.

Win/win.

Re:It's not a one way process

[Rakishi]

I on the other hand am CEO of company DEF and my software in this are is average. I on the other hand was able to pay 50% as much as the companies with "superior" products did. I likely also noticed that the costs to improve the software to be superior would actually lose me money in the long term. So I open source it. In time some competitors use it and it becomes the superior solution in the area as a result. Those competitors may have already developed their own solutions and now also had to pay the migration costs.

So now company XYZ is in a bind as it's previously superior solution is now inferior to everyoen else. It has paid more for this solution than any other company and it now has to pay even more migration costs. My company on the other hand has managed to fill in the gap in this area while spending 3/4 as much money in the process.

In the process I invested the money that I saved into parts of the company that did have large competitive advantages. At the same time I retain main control over the superior software and as a result still have a slight advantage as a result in that area.

Recession-proffing with FOSS

[SgtChaireBourne]

It may be news to a CEO, but programmers who write code (and their children) want to eat and have roofs over their heads, too.

That's the broken window falsehood in a nutshell, with a false dichotomy thrown in on the side.

Money and staff spent, in this case, re-inventing the wheel, is money and staff not spent on the core business activities. So,even if it's learning from others mistakes, going FOSS saves effort and that in turn boosts your core business activities (assuming reinvestment and not skimming by the execs). Software is only a tool, an enabler, for those core activities. In case you missed the last 25 years of computing, it's not an XOR choice between using the open source development model and making a profit. In fact, it's been show again and again that it's not only profitable, but makes your company more recession-proof. We've been through a few now and have seen the benefits.

Re:Job loss

[Metaphorically]

There's a big difference between a few implementations existing to meet different needs (or even the same needs) and the thousands of times that the same software is built for in-house solutions (because "our situation is unique"). There's plenty of good bug-tracking software out there but I know that there are also tons of Excel spreadsheets being extended to track bugs. Many of those spreadsheets have coders looking at them and telling their bosses "we should build our own issue tracking software to replace this spreadsheet."

That sort of thing happens and will keep happening. That doesn't mean that we need to have artificial barriers stopping companies from sharing code to keep those coders employed (as the GP suggested). If code is shared then there's a chance to reuse pieces or at least look at it and decide if it's suitable for another need. When it's not shared then it stays in it's buggy little corner and the same in-house coders keep tweaking it every time it breaks.

Until all code is truly reusable and free of everything non-problem-related, programmers will reinvent the wheel over and over again.
Don't hold your breath.

Hardware companies hate forward compatibility

[billybob2]

Hardware companies care only about selling a product and keeping the customer satisfied for the first 30 days after which they can't return the said hardware. They don't care if the patches to Linux don't get upstream because as long as the hardware works fine with the version of Linux that they hacked up and pre-loaded, they're customers will be temporarily satisfied.

And when it comes time to upgrade the Linux OS in a year or two, the new version won't work, so the customers will be forced to buy more "up-to-date" hardware with more hacks and band aids. Even presumably FOSS-friendly companies like System76 change the pre-installed Ubuntu on their laptops by adding tons of hacks and then don't bother to even report them upstream, much less to develop a sustainable solution.

See the following threads on the System76 forum:

Real Linux drivers for System76 laptops, NO thanks to System76

Merge System76 Driver with upstream kernel and HAL