Slashdot Mirror
Managing Open Source Projects
First Impressions There is a word for this book: SWEEEET!
It's short, too, but before you grumble about paying nearly 30 bucks for something that's less than 200 pages, you might want to look at the concept of quality. It's worth every blessed dime, plus taxes (if applicable).
Managing Open Source Projects opens with a history of the movement, thus providing background information and context to prospective or actual managers of Open Source projects. In this sense alone, Sandred has set himself apart from numerous other authors on the subject, providing an overview of a movement which has been over 30 years in the making, and whose restructuring of the "old-economy" is just beginning.
The development of the browser wars is dealt with in this history, a subject not everyone is familiar with. What it amounts to is a lesson in 'instant karma' that Netscape Inc. learned after doing a lot of damage to the Mosaic browser, ("mozilla" comes from "Mosaic killer"), and subsequently having the same done unto them by the Redmond Contingent(TM). Sandred sort of implies that this lesson had more than a cursory role in the 1998 opening of the Mozilla source code, which staggered the industry all round. Obviously, Netscape learned several somethings from the experience.
The author moves on to discuss the relevance of open source to business. (You knew *that* was coming, didn't you?). Sandred raises the common assumption of business known as Brooks' Law ('the performance of programming teams does not scale so as to increase the productivity of the team'), and then uses the history of Linux development to illustrate the inadequacy of this model in describing the open source development process. In sum, Sandred asserts that the differentiating factor is what he calls the "political attitude" of the open source model, which breeds a different leadership style. The "administrative overhead" required by each member of a development team may increase with each new developer, but without the geographic restrictions posed by a code farm, there is a wider base of "administrators" to choose from. (Now you know what to tell your boss.).
Chapters 8-10 cover a variety of tools useful (and commonly used) in building open source works, and methodologies used to set up the project (including the team). You've heard of Sourceforge, right? CVS? Or maybe "The Slashdot Effect"?
Highlights There are some portions of Managing Open Source Projects that are guaranteed "feel good" items which remind any open source developer of why we do what we do.
In his discussion of open source philosophy, Sandred points out that the viability of the open source model is not restricted to software:
"With computers, perfect copies of a digital work can easily be made, modified, and distributed by others, with no loss to the original work. Individuals interact and share informa- tion,and then react and build upon it; this is not only natural, it is also the only way for individuals to succeed in a commu- nity. In essence, the idea of open source is basic to the natural propagation of digital information among humans in a society. This is why the traditional notion of copyright does not really make sense on the Internet." (p. 52)
He points out that the United Nations has adopted an open source approach to distributed assets, including (especially) information. The link between democracy and freedom of information is clear, and iterated not only by Sandred, but by UN Secretary-General Kofi Annan, and Dr. Gro Harlem Brundtland, Secretary-General of the World Health Organization.
It's not "just" a software development model anymore.
Imperfections There are some small issues this writer has to take with Mr. Sandred's pronouncements, among them the following:
"All software cannot be developed open source. Open source software tends to concentrate on infrastructural design and back-end software. Open source benefits from incremental de- sign, which means back-end systems. End-user applications are hard to write. These applications deal with graphical user interfaces, which are very complex to write, almost always customized, and comprise other skills like graphical inter- face design." (p.160)This writer would, upon reflection, argue with pretty much everything in this paragraph, save for the self-evident last statement. Both the GNOME and KDE projects are about providing desktop applications, and the managers to go with them. Most window managers provide applications to go with their "suites". There are productivity software suites in progress.
Ah, well, it's one bad moment of two in the entire book.
Mr. Sandred makes an unwitting gaffe in his discussion of "Five Open Source Commandments" in Chapter 12: the last of these reads 'Join a project rather than starting your own.' While joining another project is helpful, even useful, it does not replace the "developer's personal itch" that Sandred quotes from Eric Raymond's 19 lessons (Cathedral and the Bazaar, O'Reilly, 1999), in Chapter 2. Do both!
Conclusion Don't just sit there -- go get the book, even if you're not currently involved in, or planning on, managing an open source project. The information is timely, the pace is lively, and Sandred has provided a wealth of insight into the open source movement's past, present and future. While some of his work has perceptual errors, these are few. The rest of it is pure gold.
You can purchase this book at Fatbrain.
Not as long the cost-benefit balance is so incredibly bad compared to closed source commercial products, like for instance this analyis shows:
Let's have a close look at the costs involved when running a Linux system.
An important factor in Linux' cost is its maintenance. Linux requires a *lot* of maintenance, work doable only by the relatively few high-paid Linux administrators that put themselves - of course willingly - at a great place in the market. Linux seems to be needing maintenance continuously, to keep it from breaking down.
Add to this the cost of loss of data. Linux' native file system, EXT2FS, is known to lose data like a firehose spouts water when the file system isn't unmounted properly. Other unix file systems are much more tolerant towards unexpected crashes. An example is the FreeBSD file system, which with soft updates enabled, performance-wise blows EXT2FS out of the water, and doesn't have the negative drawback of extreme data loss in case of a system breakdown.
The upcoming 'solution' to this, EXT3FS, is nothing more than an ugly hack to put journaling into the file system. All the drawbacks of the ancient EXT2FS file system remain in EXT3FS, for the sake of 'forward- and backward compatibility'. This is interesting, considering that the DOS heritage in the Windows 9x/ME series was considered a very bad thing by the Linux community, even though it provided what could be called one of the best examples of compatibility, ever. When it's about Linux, compatibility constraints don't seem to be that much of a problem for Linux advocates.
Back to Linux' cost. Factor in also the fact that crashes happen much more often on Linux than on other unices. On other unices, crashes usually are caused by external sources like power outages. Crashes in Linux are a regular thing, and nobody seems to know what causes them, internally. Linux advocates try to hide this fact by denying crashes ever happen. Instead, they have frequent "hardware problems".
The steep learning curve compared to about any other operating system out there is a major factor in Linux' cost. The system is a mix of features from all kinds of unices, but not one of them is implemented right. A Linux user has to live with badly coded tools which have low performance, mangle data seemingly at random and are not in line with their specification. On top of that a lot of them spit out the most childish and unprofessional messages, indicating that they were created by 14-year olds with too much time, no talent and a bad attitude.
I could go on and on and on, but the conclusion is clear. Linux is not an option for any one who seeks a professional OS with high performance, scalability, stability, adherence to standards, etc.
So hire a company to support the code for you if that's what you want. Companies like Red Hat do this sort of thing already.
I don't care if it's 90,000 hectares. That lake was not my doing.
Sure, there are OSS successes, but I can't believe the author makes a definitive argument for OSS in business as a core part of a company's business plan - there are just too few examples of real solid successes and fewer still of money-making, solvent OSS-based companies.
I'll certainly be buying a copy, as this is a subject that I find very interesting. There are a whole bunch of people out there with skills in running projects - whether they are money-generating or not - who the community could be using better. That's not to blame the community - these people have to put themselves forward, and show how they can add value to a project. They also need to understand how the community, and the community's philosophy, works. I've been flamed before for suggesting that non-coders should be allowed to be involved in OSS projects, with the suggestion that if you're not good enough to code, you're not good enough to be involved. I don't think that this is right - and there are people out there (including me) who have coded, but realise that their skills lie elsewhere, and that the way they can best further the projects that they care about is to offer the skills that they do have. It's not as if we tell the documentation teams that they shouldn't be involved!
So, how do we go about doing this? There are places where projects hang out - sourceforge, for instance - but I'm not sure that they are a good place to look for non-coders. I'd like to see projects who realise that they could use organisational help (wouldn't the code lead like to spend more time coding? Of course they would, in most cases!) looking for help - but I don't know where they should look. Functionality priorisation, tester recruitment, visibility improvement, strategic focus, project communications - all things that at least some projects could probably use.
So - how do we move forward? Maybe people like me need to make it clear how our skills may be able to help important projects, and hope that the people running them can make use of us. Maybe we need our own forum (hmm - maybe not!). I'd love to hear other suggestions.
The one quibble the reviewer has with the book seems to me likely to be one of its strongest points. Open source has not delivered good end-user applications so far; its strength has been in developer-facing products such as the kernel and the Apache web server. Open source does not meet the challenges of end-user-facing software very well because it requires non-programmer skills like interaction design and graphic design, and does not appeal much to people with those skills. In addition, good user experience design depends on extensive user testing, which requires cash outlays. Open source may produce good software by and for programmers but it falters when it comes to delivering software suitable for the average user. This suggests a hybrid model in which commercial developers create proprietary applications on top of free back-end software.
Tim
Often Mozilla is used as an example of a failed OSS project, but I don't think it is at all, both from a business and a pure 'Open Source' perspective.
From a business perspective companies such as AOL are just beginning to exploit code generated by the Mozilla project. In Spain, AOL has just launched a product called 'AOL avant', a iMac-style box allowing web browsing and email for about $15 a month, all inclusive. They are aiming to get quarter of a million of these boxes into homes. The boxes run Linux, along with a browser that I assume is based on Mozilla. When people say that Linux isn't suitable for non-techies, then this is a great example, because AOL Avant has been designed to be used by your grandma. And quarter of a million users is a lot.
It is true that the project is taking a long time, but then complex software development does take a long time. Microsoft knows this - they spend years and years refining sub-standard products until they are sweet - that's part of the reason they are in the position they are today - they don't give up. Nor should the OSS community give up on Mozilla. It is a strategicly extremely important project in the good fight against the beast from Redmond, and the team has done a great job so far.
I use it every day and it is really stable now and packed with features. For instance, it can render XML directly from XSL style sheets, something that IE 6 cannot do (or at least I cannot get it to work). Keep up the good work Mozilla crew, your work is great and your project will turn into a great success story for the OSS community!
Whoever thought Mozilla was short for "Mosaic killer" needs a history lesson.
Before Netscape was Netscape it was called Mosaic Communications. The browser they made was still called Mosaic. People may remember back in the day, pre 1.0, the betas had this spiffy logo with tiles that revolved as a progress indicator.
The browser was envisioned as the Killer Application for the Internet, much as the spreadsheet was the killer application for the PC. Somewhere in the back of my closet is a Mosaic Communications t-shirt with a big Godzilla look-alike on the back and the words "The Killer Application". That was the original notion of Mozilla. Mosaic was the killer app, we weren't trying to kill Mosaic.
Then, of course, UIUC came out and slapped us around for using the Mosaic name so the company adopted the Netscape name instead. No more Mosaic, no more revolving tiles logo. This is why we all had to live through the "throbbing N" status indicator for 1.0. They just didn't have time to come up with anything different before the release.
sclatter
Its often said that the best way to break into an open source project is to find an application you like, and add to it a feature that you would find useful.
The problem I have found with this is that few projects have any online design documents, to help someone new to the project find their way around. This is also frequently true in the commercial/closed source world of development, but in that setting you can always wander over to the lead developer's pod and ask him for some clues. Now I know you can email the maintainer, but you don't always feel comfortable taking up their time when you know they are very busy and doing this on their own time.
Why are design docs, in particular things like class diagrams, interaction diagrams, flow charts, whatever, so rarely available? Is it because the projects aren't created from design documents in the first place, or is it more a lack of a sensible way to share designs in a digital form? Are there any decent free software reverse engineering tools to create 'how it is' design documents from the sources?
<flamesuit>Is it a consequence of the kind of people heading up the individual project? I am aware that some of the most successful open source projects are run by experienced software engineers (apache being a classic example). But many of the smaller projects are run by teenagers who have recently learned to program. Is it simply that much open source software is produced by people who are not yet aware of the benefits and methods of design documentation?</flamesuit>
I'm the admin of a small open source project. What I'd like is a book that builds on the experience of others that could help me do it better. What I don't want is a history of the open source/freesoftware movement. The only reason why I'm doing is because I enjoy it, and my only goal is building a good piece of software. I don't want to change the world, and I would like to distance myself as far as possible from anyone's (IMO nutty) political movements.
Can anyone recommend anything? The review suggests this book is not what I'm looking for.
I've had the opportuntity to work with companies that are either using open source software as a backbone for their services and for their products.
An area that I've seen in which all have many challenges with, is not in the development of their respective products and services, but in managing the environments, tools and source code.
This is not unique to open source, but this is an area where open source could really contribute. Fundamentally the tools used by the software industry to manage and publish software are 20-30 years old and have not undergone much in the way of evolution. The tools and practices used today to manage the development of software, in most cases gets in the way of productivity and quality, because they are not DESIGNED to solve todays problems, they are adapted to solve today SYMPTOMS.
If we look at all of the mainstream change control systems, they Generally don't solve todays business needs, resulting in customizations by the development teams, that generally miss the mark. They may support some engineering team needs, or some developer needs... But ask most team leads, project managers, and engineering managers and they will tell you that this is an area that causes alot of pain.
CVS based on top of RCS
Teamware based on top of SCCS
Perforce based on top RCS
Continuus based on top of RCS
ClearCase based on Apollo DSEE
Bitkeeper based on SCCS
SCCS was developed... what around 1975, RCS early 80's, ClearCase/DSEE mid 80's
Fundamentally these products are not DESIGNED to support all the different variants and product lifecycles to repsond to todays business opportunties.
Individuals and Companies have made great attempts at creating wrappers, customizations, books, and articles to attack the problems that the deisign and implementation of these solutions failed to deliver. But this is only addressing the symptoms of the design and implementation of these tools
A new look at the problem is needed, if anyone expects to see any significant improvement in the ability to deliver the right solution in the right timeframe to the right customer.
I'm not sure, based only on my exepriences up-to-this-date, that open source methodologies can make any SIGNIFICANT improvement over todays current business methodologies... I don't see any accountability in open source metholodgies. Not being able to see accountability in the methodologies, leads me to be concerned about how open source addresss businesses, support the consumers need to effectively resolve issues and concerns.
Regards,Kramer
This review didn't tell me much more than the table of contents would have. It would be nice to know what tools the author describes, the kind of examples that are included, and a general sense of the technical level.
I say this because I am in charge of a large project. I've never managed a project before. Most of my learning comes from trying new stuff out and learning from my success and failure. I don't have the luxury of a mentor, and I don't have the luxury of a technical bookstore (one of the perks of living in a very small town).
Something other than "this book rocks, huh huh" would be nice.
Anyone else have a more detailed description of this book?
The middle mind speaks!
The secret, boys and girls, is a web-based UI. Granted, that doesn't fill the bill for a lot of applications, but for a pretty good number of applications it works just fine.
It's called constraining the solution space.
This really means, "don't re-invent the wheel". If your project goal is to get working software package, it is quicker, easier, and hopefully has fewer bugs to work with and existing project and try to achieve synergy than starting another similar oriented project from stratch.
See: Never do this
If you just want to write your own text editor that is fine too, but the benefit isn't necessary as great as what is possible through colloration.
To those of you who wrote, thank you for the references. One of the main weaknesses in the whole software development publishing business is a lack of books that fills the gap between coding and planning. There are lots of books that describe coding techniques, Effective C++ comes to mind as an example of this. There are also lots of books on high level design, like Design Patterns, and management, like the Extreme Programming series. However, there are very few books that adequetly describe the integration of code, source control, development environments, and makefile management, and document development (including things like autoconf, automake, cvs, etc.). It's that integration of all those things that continues to bite me in the ass, and I feel like a blind man making my way through a unfamiliar kitchen with lots of sharp knives anywhere. Sure, books like The Pragmatic Programmer say things like "source control is good" and "automated testing is da bomb," but don't give any concrete examples to get one started down the path of effectively integrating those tools.
Once again, thanks for your suggestions. Some of them may prove very valuable (I've ordered a few books and checked out some references already).
The middle mind speaks!