Forums : Posts

Better detection of duplicate projects is on our bug list. For now, they are uncommon enough that we just clean them up manually. I've deleted 4488, since it appeared to be the lesser of the two pages.

Hi fsck, The summary is completely automated through source code analysis. The only way to change the summary is to change the source code. For basic metrics like number of contributors, number of ... [More] commits, and number of comments, each project is compared to the typical values for other projects on Ohloh. Projects which fall outside the average range are highlighted with bullet points in the summary. Raw source code metrics can be very dry and detailed. The summary is our attempt to highlight some big-picture facts about a project in a way that's easy for the layperson to understand. [Less]

Oops. It was all my fault. I recently made some performance enhancements on our project pages. The new Links query had a bug - which is now fixed. Thanks for your alertness and patience.

It'd be great if the most recent figures approximatively displayed in the codebase history graph (code, comments, blank) were made explicit, e.g. code.............2123 lines.....64.5% ... [More] comments.....1032 lines.....31.3% blank..............134 lines.....4.2 % this would be especially useful to understand how specific projects position themselves with respect to the reference values Ohloh uses to establish factoids. For example: projects are marked as well documented if their comment/code ratio (N%) is above the average (M%) for that language category. At the moment, though, there is no way to know the comment/code ratio for projects not highlighted through factoids. [Less]

Hi all, Great stuff - I'm loving ohloh and added several projects and built my stack. Your stats helped me pull together some stats for a presentation I am giving. I'm so glad I didn't need to ... [More] manually go through some changelogs in source code to figure this stuff out. :-] Any plans to add some aggregated stats for someones stack? I'd love to see, for example, total LOC and person days for all projects in my stack. Keep up the good work. I've been happy with support so far too. Thanks [Less]

Hi Robin I guess this is a known issue at our end here - fails for me on Mac too. Depending on what O/S your SVN client runs on, it may fail on at least one file that is encoded poorly. Looks like ... [More] there are some workarounds if you can set certain environment settings, but I assume this isn't an option? Who knows, maybe it is a non-issue for your client - I'll cross my fingers :) I also gave a bad URL for the SVN source which included a bunch of website content (ugh) So I've added a second enlistment that only targets the source code subdir and doesn't include a bunch of website content that is likely binary. Perhaps we can kill the first enlistment and see if 2nd goes faster?? [Less]

Yes, that's true. If your project does much of its development work outside of the source control system, then Ohloh will not see that effort. That's why I'm a growing fan of source control systems ... [More] like Git and Darcs, which preserve the history of all work done anywhere, even if the developers don't have commit permissions on the master repository. When outside work is eventually merged into the master source code, the entire history of the outside code comes with it. [Less]

I asked this via email also before the forums existed. A result of their methodology is that it heavily skews the statistics for CMS tools; webservers; and databases that put a link to their own site in each comment or page posted.

Ok, thanks a lot. I also had the same issue while working on a script to automate some SVN operations. I don't know if the auto-accept-certificate feature is intentionnaly not implemented by SVN ... [More] devs. If not, it could be a good idea to submit them a feature request. [Less]

Hey I really like the graph that shows for each contributor how active they have been. However, its on a very large scale - 5 years. Can you base that on how long the project exists? If you go to the ... [More] Code tab, you'll find: 'The source code repositories show that the project is x months old.' Where x is a number of course. Why not use that number to base the graph on. Lets say my project is about 11 months old, then each contributor would have the graph reach to lets say 1 year. And once it becomes 12 months old, the graph changes to 1 year and a month. That would make it a bit clearer I think, however, I do like how it is now. Thanks [Less]