Hacker News Reader

Comment by davidw | Hacker News Reader

My numbers are better than theirs: http://langpop.com - I use multiple sources. While I stand by my results, I think that they are, by their very nature, not all that precise, so making a big deal out of a blip like this is great for publicity, but probably not very meaningful.

View on HN

Consider using log charts to show the differences at the lower end more precisely. This would also save screen space: you could make the bars vertical rather than horizontal.

not sure if freshmeat and google code are the all that 'authorative' about the opensource projects using languages. for one most Ruby projects are hosted on github.

I'm more interested in releases than 'hosted projects', as many of the latter are unfinished. Which is why I've always used Freshmeat rather than Sourceforge.

I agree that it's probably a bit long in the tooth, and I'd be happy to add something newer, as long as it's not too terribly biased. Which, however, I think github (and StackOverflow) continue to be: the former in favor of Ruby, the latter in favor of various Microsoft things.

well, the best thing might be to just combine all the project listings you can pull from all the sites together. for what it worth you can try to only pull recent projects, or projects passing some minimal threshold of 'usage' like download count on sf or watch/fork count on github. but w/o counting the github you are severely undercounting ruby projects. the ones on google code and freshmeat are a statistical aberration to the real count :) ruby people just don't care enough about freashmeat to use it much. all that matters is really just on the github. 'releases' announced on blogs/mailing lists and by releasing new version of a 'gem'. no need to duplicate it on some horrible 'releases' site still stuck in the 90-s :). rubygems package manager does a great job of informing people of new versions.

Thanks for the additional data.

If you don't mind a suggestion, having the combined index at the top, with details below, would probably be more attractive.

The large logo with vertical color bars also creates an expectation that the monochromatic horizontal bars end up disappointing ...

I strongly agree that the normalized/aggregated results should be at the top. The first graph confused the hell out of me (like erlanger, I thought "D? Seriously?").

Think of it as putting an executive summary at the start of an article or paper.

For all this bluster, D seems awfully high on that list (above ActionScript and Objective-C? really?).

I guess ableal is right. You need to be looking at the normalized results, further down the page.

Ah, that makes sense, I didn't read and assumed that the first chart was the summary.

Reply on news.ycombinator.com