More MS, Less Talent In Open Source's Future

alphadogg writes "The open source industry in 2008 will be marked by more news out of Microsoft, IBM, Oracle and other big IT vendors, less start-up funding, more M&A activity, and an increasingly serious talent shortage, according to Raven Zachary, open source research director for The 451 Group. One example of the talent shortage will be people with expertise in the Tomcat open source Java servlet middleware from the Apache Foundation. 'There are 25 or so core contributors to that project,' Zachary said. 'Over the past four or five years that number has stayed virtually [unchanged]... but the growth of Tomcat has been astronomical.'"

Talent shortage?

[RandoX (828285)]

Maybe talented coders like to get paid better.

Re:Talent shortage?

[wwmedia (950346)]

u mean not all programmers like to give away their work and answer support questions for free with their open source software?

Re:

[wizardforce (1005805)]

there's nothing in gnu saying you can't indirectly make money from software, you just can't redistribute code that is derived from gnu code in a propietary way. Canonical I hear makes about 50 million a year through support and indirect revenue sources from Ubuntu

Re:Talent shortage?

[by Anonymous Coward]

Exactly. Unfortunately, support and consulting are arguably the least desirable way to make money in this sort of industry. Ideally, you'd sell product and get only positive feedback to improve it. No support. No dumb questions. No issues.

Companies that make money from support contracts are, in my opinion, doing the least favorable work. It's certainly not sexy and for every dollar you earn, you have to work an amount directly proportional to that. There's not much concept of exponential growth. In other words, your income per hour flattens out much faster than with a product-based model.

Re:

[yason (249474)]

Companies that make money from support contracts are, in my opinion, doing the least favorable work. It's certainly not sexy and for every dollar you earn, you have to work an amount directly proportional to that. There's not much concept of exponential growth. In other words, your income per hour flattens out much faster than with a product-based model.

It pays to do work that nobody else wants to do.

Further, I don't think there will be "exponential growth" (or scalable sales...) in the future softwar

Re:

[eggnoglatte (1047660)]

This must be the first time I have heard somebody call the MS business model unsustainable. Sure, it isn't very nice, but unsustainable? How exactly did they end up in that dominant market position for close to the last 2 deceades, then?

Only one way.

[Benanov (583592)]

>This must be the first time I have heard somebody call the MS business model unsustainable. Sure, it isn't very nice, but unsustainable? How exactly did they end up in that dominant market position for close to the last 2 deceades, then?

Illegally.

(Remember? Convicted Monopolist.)

Re:

[Reverend528 (585549)]

there's nothing in gnu saying you can't indirectly make money from software

There's nothing in the GPL saying you can't directly make money from software.

Re:Talent shortage?

[kdemetter (965669)]

so don't use GLPv3 . There are plenty of other open source licenses . It's not because the FSF makes a new one that you have to use that. There is no good or bad license in there , it just depends on your needs . I don't think there's a talent shortage , it's more like a shortage of new ideas . Basically , if you think of an application you could use , there's a good chance it already exists . And it's a lot more fun working on something you helped create , than on improving something someone else wrote .

Re:

[by Anonymous Coward]

if they use that code to hinder the modification of modified gnu source code than yes GPLv3 would prevent that, if their code was a seperate entity that in no way affected the modification of modified gnu code then there is by design, no problem. Tivo allowed you to view the source but not to modify it for your own use on their hardware, that would be considered anti-ethical to software freedom. Then again there's nothing preventing said company from using BSD code or writing their own code for the job no

Re:Talent shortage?

[RHSC (1019802)]

I prefer to get paid by the semicolon...just so long as I don't have to code in VB

Re:

[morgan_greywolf (835522)]

I like to get paid by leading whitespace -- syou know, paces and tabs (with tabs counting as at least 4 spaces). But then again, I'm a Python programmer.

Re:

[RHSC (1019802)]

Getting paid by the parenthesis while coding LISP is good too

Re:

[MenTaLguY (5483)]

Absolutely. But Open Source gives you the freedom to do things the way you want to, and it's a good way to interact with a population of other talented coders, both of which are generally hard to get consistently in the commercial world. Both are things that talented coders care about a lot more than the folks who are just in it for the money, clocking in the hours and punching in the first code that dribbles out of their brain.

It's more a question of having the spare time for it.

Re:

[Reverend528 (585549)]

Maybe the talented coders just don't like Java. It's entirely possibly that a shortage of tomcat coders will correspond to a surplus of rails developers.

Google trends seems to agree with this theory [google.com]

Re:

[PrescriptionWarning (932687)]

not all open source is derived from hobby work ya know... some of it is designed by programmers of 'your' generation because it was a part of what they did for pay.

Re:Talent shortage?

[xanadu-xtroot.com (450073)]

not all open source is derived from hobby work ya know

Perhaps, but you're missing his point. Here's another example:

• Wake up around 6 am. Eat, get dressed, and get physched up for the fun fun hour plus drive via major congested roads to get to work by 9am. That pre-leave time also includes helping get my Son ready for School (breakfast, persistant reminders that he's running late, etc.).
• Work 'till 5pm (usually closer to 6 pm).
• Fun fun hour plus drive home. Home around 7 pm.
• Eat dinner, help Son with homework, spend a bit of time with him before his bed time.
• Spend sometime with Wife and un-wind a bit (and "un-wind", if you catch my drift...)
• Shower, go over meeting notes, maybe catch the the earlier late news).
• 10pm Head to bed.

Where in there am I supposed to find time to sit in front of my machine spending hours debugging code for an OSS project? I'm not saying that I don't contribute, once in a while I have sometime on weekends to submit a bug report (with some same code usually - but not always), or something small like that, but by far and large, us "older" (I'm only 34, but...), "Family Guys" simply don't have the time the younger people (in High School or College) do.

That was his point, I think.

Re:Talent shortage?

[dvice_null (981029)]

> Where in there am I supposed to find time to sit in front of my machine spending hours debugging code for an OSS project?

"We" have the same amount of time what everyone else has. It is just how we want to spent it.

For example I moved to very close to my current work place when I started working there. It takes about 10 minutes for me to get to work and I don't even have to use a car for that. I save probably 10 hours every week compared to you. That is something like 500 hours every year (+ I save a lot of environment and money at the same time).

I like to think that my skills are too valuable to be wasted in traffic jams every day. I would ever turn down a job, if I couldn't move close enough to it. You obviously have different priorities, which are probably better than mine. But you really can't claim that you would have less time than anyone else. (I also have a wife, child, job and I spent my free time on Slashdot and with open source projects.)

Quantity != Quality

[dintech (998802)]

Tomcat is an excellent product and a gem of the open source community. Just because there are 'only' 25 core developers working on it doesn't make it inferior in any of the other offerings out there. I'm not sure throwing more developers at it would necessarily make it better. See, Mythical Man Month for details...

Re:

[by Anonymous Coward]

I beg to differ.

I tried to install Tomcat the other day for a rehosting consulting job I was tasked with.

The initial part of the install went fine, though the documentation seems to be written by someone from another planet. Very strange verb tenses, grammar, poor train of thought throughout (very jumpy).

Anyway, after I got Tomcat up and running, I realized I needed a connector to hook it into Apache. The docs were kind of sketchy on this (yes, they brought it up, but not in an organized, linear manner. It'

How is this a troll?

[Pengo (28814)]

I actually agree with the AC.

I've been running Caucho's resin as an alternative to Tomcat for many years and it's been an outstanding product with none of the headaches I've had with Tomcat. Resin is GPL'd with very good documentation and optional low-cost commercial support.

Just because it's from apache doesn't mean it's the best for the job at hand. I find more often than not, people use tomcat because they believe that there are few options available, let alone easier and more elegant open source solut

Re:

[LDoggg_ (659725)]

The AC's post wasn't a troll, the moderator was trigger happy or misunderstood his point.

He was right, tomcat used to be a complete pain in the ass to connect to apache web server. Thankfully things have gotten much easier.
Also, on Fedora 8 you can have this all automatically working with the new open source JDK.

Re:

[LDoggg_ (659725)]

As of apache 2.2(web server, not tomcat) mod_jk is obsolete and this has gotten a whole lot easier. Take a look at mod_proxy_ajp.

It's now just one simple proxy_ajp.conf file. Plenty of options for advanced configuration, but a simple configuration could be done in one line like "ProxyPass /examples/ ajp://localhost:8009/jsp-examples/"

What you need is *not* a *core* developper.

[DrYak (748999)]

That's the typical opensource situation where whom you need is NOT a core developer.

25 developers are a pretty good team to constantly write, re-write and improve the inner workings of tom-cat. In fact, there are a lot of commercial project that don't have that much developer 100% dedicated to the project. And as GP poster pointed out : "Mythical Man-month" explains us why this team doesn't need to grow much more because of the added inter-communication and training of newcomers overhead.

What a lot of newcomers into the OSS world fail to realise, is that there is a lot beside "writing code" that is important for an OSS project to be useful. There's, for example, a very strong need for artist to make the visuals (UI design, themes, other graphics) in order to avoid having the OSS project look like some 10 year old ass-ugly Athena interface with a cryptic UI based on a non obvious metaphor.
And, like in your case, projects also needs people with good writing skills, to write nice documentation, specification, HOW-TOs, and other guides, because frankly there are a lot of OSS projects out there that are technical marvel from a technological point of view but whose documentation consist mainly of a a big dump of code comments and function names and where, in fine, the old classic formula "Google + {error message} = posts in newsgroups" is the only way to get decent help.

People usually fail to realise it. For them Open-Source mostly remind them of complex C/C++-code and they think that GPL is only for programmer good at writing code. And thus a lot of people aren't motivated to contact a project and start helping because they think they don't have the necessary coding skills. Whereas in fact, even with no competences at all in programming, they could be critically important with their artistic, litteracy, or other skills. (Even things like helping organising appearances of the project at major Meetings and Expo can help because it bring attention to the project, and that requires skill that are neither coding nor artistic).

Why?

[Gotung (571984)]

Why does an open source project magically need more programmers because it has become popular? What's wrong with the 25 guys that have obviously been doing a kick-ass job with Tomcat? Throwing more bodies at it will just lead to bloatware.

Re:Why?

[ByOhTek (1181381)]

I think they were stating it's size and complexity as the rational, not the popularity.

Still, yes there are 25 core contributors to Tomcat, but what is the total contributor size in a per-mont/per-year breakdown for the server.

And what percentage of the updates are being done by the core developers? If the proporition of the development done by the core team is half of what it was the year before, at any given point, but about the same absolute amount of work - then the development on the project is still growing exponentially, even if the core team remains the same size.

Re:

[cromar (1103585)]

25 developers is a lot. Maybe it's just me, but 25 developers would seem to be enough for most large projects. The real question is "how many developers dos it take to screw in a light bulb?"

Re:

[dekemoose (699264)]

This is true. I was actually fairly surprised to find out the number of developers that my company actually has (or had before we were acquired). A few good developers is all that any good project/product seems to need. Adding more seems to simply result in more problems, not more progress.

Re:

[plague3106 (71849)]

One; to call fascilities and have them come out and do it.

Good use of resources?

[MikeFM (12491)]

If you can accomplish great things with a few core programmers that is called being effecient. Adding more programmers to a project usually makes it worse rather than better. Open source allows many developers to make minor changes, as they have need to, but doesn't change the fact that only a few core programmers are needed for most projects.

I don't see the number of open source programmers shrinking at all. If anything, I expect to see many new projects taking shape and a few catching fire and shaking up the industry. It's better for many small projects to be seeded so that a few can grow into new major projects. There'd be no point in adding more and more developers to existing projects.

Growth of developers vs. growth of users

[Iphtashu Fitz (263795)]

Just because one of these groups increases in size doesn't necessarily mean the other one has to. I've worked for fairly small companies where the number of developers didn't change dramatically despite the rapid growth in end users. Sure, more developers may be hired if you start developing new products, dramatically increase the feature base of the existing product, etc. but for projects that are relatively stable and have slower growth cycles there really isn't a need for a growth in the number of dev

From TFA

[Silver Sloth (770927)]

"Microsoft is still trying to work out its strategy," he said. "Ultimately, I think we'll see them embrace open source much more."

Now I'm worried.

Re:

[Yetihehe (971185)]

Don't worry. I think open source is bigger than MS and when MS will try to embrace it, it will expand and extinguish MS.

Number of maintainers falls as project ages

[xant (99438)]

High-quality products general stay flat or lose developers over time without losing any quality. I have no idea whether tomcat is a high-quality product or not, but the core of it probably requires very little maintenance now, leaving the "core" developer circle free to work on edge features. There are an unlimited number of those for any given project, but the urgency of those edges falls off rapidly as a project ages, so it's rarely the case that a project needs to grow in developers just because it's getting older. Such projects usually split into separate projects with their own functionality core.

Also, it's ridiculous to extrapolate this process and make a statement about all open source. Developers are rarely destroyed, converting their energy into entropy. Instead, they are simply attracted to new products that need developers.

Finally, the talented open source developers pool will only grow, as it always has. If Microsoft is hiring people to work on open source, then those people will be new talented open source developers.

Re:

[cnettel (836611)]

A "product" like Tomcat is useful by being a platform. As such, new standards and other developments outside of the product makes a greater implact relative to, say, a game. Many Java-related projects (i.e. projects that not only are written in Java, but support other stuff in Java) went through significant changes for the 1.5 release, for example. Just keeping a project of this type and quality alive requires quite a bit of work.

I don't see it

[Wylfing (144940)]

First: who the F cares about announcements from Microsoft regarding open source projects, unless they are actually contributing.

OK, that out of the way, I can't see how a shortage in one project is a shortage overall. OS is about coders scratching an itch. I have contributed to projects but only when it was something that impacted me personally, and I wanted to see it fixed in a hurry. If the number of users of a project grows astronomically, that's great, but it has no bearing on how many coders participate if nobody feels an "itch" they need to scratch. Maybe the software is good enough for end users, and they feel fine about it.

Those coders aren't "gone." They're just off scratching some other itch, is all.

Huh?

[snoyberg (787126)]

'There are 25 or so core contributors to that project,' Zachary said. 'Over the past four or five years that number has stayed virtually [unchanged]... but the growth of Tomcat has been astronomical.'"

I don't get it. There's an open source project run by 25 or so people that's had "astronomical" growth, but since they aren't bringing in new people there's a lack of talent? If they're doing well with those 25, why does the team have to grow?

25 is about 15 too many.

[Organic Brain Damage (863655)]

The assumption that a bigger team is an indicator of health is insane. Large teams in software development spend most of their time NOT WRITING CODE and NOT DEBUGGING CODE. They spend their time in meetings trying to figure out how to get 25 people or 50 people to all work together. If you have a really big job, like making a modern spreadsheet product, your best bet is to figure out how to partition it into a series of jobs that can be handled more or less independently by separate 5 person teams.

Re:

[Wylfing (144940)]

If they're doing well with those 25, why does the team have to grow?

Precisely. Raw numbers of coders don't mean anything. And now for the real reason for my reply...

In Soviet Slashdot, a beowulf cluster of Natalie Portman imagines you

Oh, please let that be true.

Re:Huh?

[Shadowlore (10860)]

Because that is how management thinks. I'll break It out for you.

Management types want more people to manage because it gives them a means to argue they deserve more money. Their management wants to see more money first. SO if your product is successful and growing, your management expects to be able to bring on more workers so they can be considered more important and worth more money. Think of it as HR bloat just like feature bloat in an application.

Since these "analysis" articles are done by people who are trained in, experienced in, or familiar with that model that is what they expect of everything. It's the notion that success brings growth. They are blissfully ignorant of the small world concepts, or how real work gets done, or how software is different from building a Model T, and only see the "business" side - especially since that is what pays their salary.

Okay, time for the car anology

[zappepcs (820751)]

How many automobile manufacturers were there at the beginning of the 20th century? How many do we have now? Since the number of manufacturers has changed little at all, or even shrunk, can we assume that even with astronomical growth in the use of automobiles, that there is something wrong with the automotive industry?

While that doesn't quite fit perfectly, I think it makes a point. If your 25 coders are putting out code good enough for astronomical use growth, then no more coders are needed. Every OSS project does NOT have to turn into a MS look alike to be successful. I think the author needs to re-evaluate their definition of success here. The hummer vehicles are successful as business goes, but there is not one in every driveway in North America yet. I have some very successful code, and there are 3 users total. It hums along nicely, 24/7 doing it's thing and all the end users are happy. It does not have astronomical growth, but it is SUCCESSFUL.

Why does F/OSS HAVE to compete with MS? That's not really rhetorical. For most of what I do, OO is absolutely great. I have no need to run and load MS Office. To me, OO is successful. I don't have to drive a Silver Ghost to have a great car. Tomcat and Apache are very successful at what they do because (IMO) MS sucked at that job and offered no real competition.

MP3 players are a successful market... not because of the superior sound quality, or because they were made by MS, but because they do their intended job very well. Some better than others, but all do the job. In the software world, it seems rare that there are more than two options for a given product precisely because of MS (not counting Mac products). If you only had a choice between an H1 hummer and a Mitsubishi Galant, or a BMW motorocycle... which would you drive?

The insistence that software must be like MS is at best absurd, and at worse, it's the worst thing that could happen to the F/OSS software industry.

Got served a WebSphere ad with this story

[Graftweed (742763)]

Is /. using targeted ads now? I was reading this story using the RSS feed and the annoying embedded ad image proudly (and weirdly) announced:

"Using Tomcat but need to do more? Discover WebSphere Application Server."

Tomcat?

[etnu (957152)]

People who use Tomcat tend to be enterprisey types (which perhaps goes without saying; using Java to solve web problems is like using a chainsaw to shave), so it's no surprise that few of them are willing / able to contribute to the project. The kind of domain knowledge required to create an http server and to do the wiring necessary to make things easily configurable is pretty far removed from the typical day to day work of these engineers. I think this is kind of true for most open source projects. The f

Network World FTL!

[Jonboy X (319895)]

He also said he anticipates Microsoft becoming increasingly busy in open source, since it "has a vested interest in making sure open source works well on Windows." However, he noted it could be well into the next decade before we see something as dramatic as an actual Linux distribution from Microsoft.

NetworkWorld: Your source for alarmist headlines, buzzword-compliant articles and wild speculation for over 20 years [networkworld.com]

Re:

[alphadogg (971356)]

Gosh, thanks for the new tagline. BobB, Network World

Boomer retirement will increase open-source devs

[gregor-e (136142)]

The much-predicted talent shortage arising from the retirement of the baby boomers may, paradoxically, swell the ranks of open-source coders. Open-source folks work for coin of the spirit (which is that same thing wage-slaves work for, ultimately, as they turn their coin of the realm into stuff they like). Anyway, retired folks get itchy for something to do, and no longer need to earn a living. A lot of them will still have viable coding skills, and I expect we'll see a groundswell in open-source develop

Anyone want to give me a job then?

[xaxa (988988)]

In July I'll graduate with a 2.1 (or better!) masters degree in Computing from Imperial College London (which, in case you don't know, is ranked very highly for computer science). I'm currently looking for jobs based in Europe, preferably in a very large city, preferably London; free software would be excellent! :-D

Email me :-). (Or email my University! rsi at doc.ic.ac.uk is the careers person. Not enough interesting companies do this, but all the investment banks etc do -- I get at least one email a d

Re:MS...

[teknopurge (199509)]

Chairs.

Everyone knows that.

Re:

[wwmedia (950346)]

with an alexa rank of 152 and involved in running a site of comparable size that uses adsense i estimate sourceforge makes 2,000 -> 3,000 $ a day from adsense

Re:

[JamesP (688957)]

Fact: Programmers need money to survive and are generally underpaid.
Fact: People can work only 40-60 hours a week without burning out and writing crap code.
Fact: Programmers have lives outside of the code.

For Open Source to survive, it's going to have to figure out how to compete in a market economy.
Part of that means making better code, since some OSS projects (OpenOffice) are total garbage full of bugs.
Part of it means a path by which the average OSS application can monetize itself and pay its developers.

Maybe SourceForge needs to distribute profit from its AdSense earnings, I dunno.

Funny...

Most places I see the kind of problems these 'facts' show are closed-source shops.

Oh yeah, another 'fact' for you. Open Source projects kicks closed-source projects in the groin in software best practices, construction techniques, usage of tools, etc, etc