It looks like there was actually one failed job when importing the hundreds of Eclipse modules. I resubmitted the job, and hopefully it will succeed and the report will be available today. If not, it ... [More] will probably require some investigation. Thanks, Robin [Less]

Whenever I try to issue a cvs 'ls' command against the US OpenBSD hosts from one of our servers, I get a "failed: Connection refused" error. European servers are timing out. This will probably take ... [More] some investigation. I'll look into it further when I return from vacation after the holiday. Thanks, Robin [Less]

Sorry, donert. An Ohloh report simply can't be generated without a revision history log. We get a lot of requests to support tarball snapshots, and as time goes on and development resources become ... [More] available we may come up with something here. Although we couldn't make very detailed reports, there are some usable metrics that could be derived. Jason wrote a good post about this recently in the blog. [Less]

I wrote a script to import all of the Eclipse CVS modules into our database. It's running now. Curiously, the dev.eclipse.org CVS server doesn't like the 'ls' command -- each attempt ends with the ... [More] error 'Terminated with fatal signal 11'. Since the 'ls' command is used by our validator, this is something of a problem for us. I'm not sure if the problem lies in the cvsnt client we are using or on the Eclipse server. I worked around the problem for purposes of the scripted insert, but note that you won't be able to add Eclipse modules manually thorugh the website because of this problem. I'll keep investigating. Robin [Less]

Sorry, Davis, I got distracted. I started a fresh download of the project. Thanks, Robin

entel, Support for more languages is always on our to-do list because there are a seemingly infinite number of them. However, we have been hearing about erlang from many people, and there are ... [More] evidently a lot of projects on Ohloh that use it. I'll push for erlang support, although it will probably take us until after the new year to get it going. BTW, do you happen to know if the source control repository for the erlang language itself is available? It would be nice to add erlang as a project on Ohloh. Thanks, Robin [Less]

entel, I agree that "Small development team" is a mischaracterization of ejabberd. The factoid is technically true, but it discounts the work going on behind the scenes. As you already understand ... [More] , Ohloh derives all of it reports strictly by reading the source code repository logs. This is a mixed blessing. On the positive side, it allows us to render some very basic, undeniable truths about a project which are impossible to find elsewhere, and which are very difficult to falsify. On the down side, it can make us blind to things which do not happen directly in the source code. It may not console you much, but all projects on Ohloh are subject to the same mechanistic rules. I readily agree with your proposal #1. We will take a look at the wording on the project page and in the factoid and try to come up with something more precise. BTW, if you are using a development model where a few administrators are responsible for committing patches from numerous developers who don't have write access to the repository, you might want to look into a source control system that better matches that model. Git (and I suspect Darcs) will do this, and this is the source control system and patch administration model used by the Linux kernel team. When an administrator approves and applies a patch to the main repository, that patch preserves its authorship information, which is why the Linux kernel project report on Ohloh lists over 1000 contributors. Another nice feature of Git is that you don't need a central server to host the code. Can you tell that I am a big fan of Git? Proposal #2 begins to address more directly the main problem we are discussing: what happens if someone disputes aspects of the Ohloh metrics, or wants to add some commentary? For now, we've put a discussion forum on each project as a place to discuss the report, but I admit that these forums are not readily visible and there has not been much activity on them. This is a general question we're still thinking about. Ideas? I like your proposal #3 very much as well. We would additionally like to link to the LICENSE and README files. There are some technical hurdles for us to overcome before we start allowing online browsing of the source code, but that is a feature we are planning to implement. We want Ohloh to be simple, accurate, and trustworthy. I welcome your input. Thanks, Robin [Less]

Yes, msevior, I agree with mokhov and hope that this helped you. If you are still having trouble please tell me exactly what error you were seeing and I'll be glad to help. Thanks, Robin

Hi Davis, The new report is finally ready, and we've got the surprising result of negative lines of code during the early history of the project. Is this something you might have expected based on ... [More] how you edited the repository? Should we dump the report completely and start over with a fresh project? Robin [Less]

I checked the LOC by hand on a local enlistment, and the LOC is correct. I suspect you've added much more code to the Ohloh project page than you intended. The source code URL entered for this ... [More] project was http://svn1.cvsdude.com/saladyears/Cities3D/trunk, which includes the compete source code and documentation for a lot of third party libraries: boost, wxWidgets, fmod, etc. I know that boost alone has over one million lines of code. If you want to see just the contents of the Cities3D directory, you can undo this source code enlistment and replace it with a link to http://svn1.cvsdude.com/saladyears/Cities3D/trunk/Cities3D. [Less]