source code?

In which language is cluto implemented? Is the source code available?

RE: .


RE: CLUTO is written in C, and

CLUTO is written in C, and the source code is not available.

RE: Will it ever become open source?

It has been 3 years since CLUTO has been updated ... will it *ever* become open source? I would like it to be, and here's why:

-- Software typically has one bug per 2KLOC, depending on the project history. I assume that this means that CLUTO probably has dozens of unpatched bugs, and (like many folks) I am unwilling to make a large time investment in s/w that cannot be fixed when a show-stopper bug is discovered. (There's also a significant risk that one of these bugs is a security bug, which means that deploying CLUTO in any sort of web or server environment is dangerous).

-- Software that hasn't changed in a long time becomes harder and harder to install on new systems, typically because newer libraries and systems are no longer compatible with the core assumptions made by the old s/w. Although these kinds of compatibility problems are usually trivial to fix, the lack of source/support means that they are unfixable. At some point, the burden becomes overwhelming (in my experience, this happens for s/w binaries between 3-7 years old).

-- Users always have new requirements, new needs, new ideas and new uses. For software to stay vital, it needs to keep up with these. Software that doesn't keep up becomes progressively less usable and less pleasant to use (as compared to the alternatives) -- and again, in my experience, this bit-rot starts happening in the 3-7 year time frame.

-- The economic advantage gained from keeping good source code closed is usually illusory, and outweighed by the advantage of opening it. So for example: 3 years has been long enough to resolve the idea that a profitable start-up might be built around CLUTO -- the original author will not actually strike it rich by licensing the code to some imagined business. By contrast, there is a real economic cost in lost reputation: the original author cannot claim that their s/w is used "by everyone", that it sets a "world record", that it is a "fundamental cornerstone of clustering and data analysis", that it "sets the gold standard". Building up this kind of reputation is important for tenure considerations, for promotion, for demonstrating leadership in technical vitality, for getting considered for, and getting accepted to advisory boards and committees, for being offered chairmanship, and even for getting one's students a good postdoc or even a good job in the industry, etc. In my experience, the additional economic advantages gained by reputation far exceed any losses in license fees from closed-source s/w.

That's my basic argument. I hope that at least the core CLUTO becomes open source soon.

Dr. Linas Vepstas