FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister 
 ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 

Project Activity Stats

 
Post new topic   Reply to topic     Forum Index -> Site
View previous topic :: View next topic  
Author Message
brad
Site Admin


Joined: 22 Feb 2004
Posts: 490
Location: Atlanta, GA USA

PostPosted: Tue Mar 21, 2006 1:02 pm    Post subject: Project Activity Stats Reply with quote

Project Admins,

I've started a new ticket for this topic, and would like to discuss here in the forums. Then I'll distill down thoughts and make them comments in the ticket. Feel free to comment in the ticket as well, but healthy debate should be here.

Here's the deal: I want to get these metrics per project (per XXX timeframe) and have an overall project activity index be available on the /projects page so visitors may sort by activity level, and ID the dormant projects from the active and/or relevant ones.

Here's what I was thinking so far (copied from ticket):
Code:

 * number of commits
 * last commit date
 * status
 * average size of commit
 * SLOC
 * downloads count
 * repos checkouts
 * trac_env page hits
 * site page hits
 * wiki pages (count, last edit) ??

Is there some sort of formula and/or weighting system that could be used to fairly represent a project's status? Could we borrow from sf.net's algorithm?

If a project is 5 - Production/stable, and hasn't had a commit or anything in a while, that shouldn't penalize them, because it could "just work" - it should be reflected in downloads, etc., but I'm not sure how to put all of this together. Also, I have to get the metrics first.

Should there be a time grain to this? i.e. each of these stats kept monthly or weekly, over time?


Thanks,
Brad
Back to top
View user's profile Send private message
pragma



Joined: 28 May 2004
Posts: 607
Location: Washington, DC

PostPosted: Tue Mar 21, 2006 1:46 pm    Post subject: Reply with quote

Neat idea. I have a few comments to help simplify things:

This all reminds me of a similar protocol used for the expiration of passwords in a system. So a system would rate projects into the following categories:

1) Active - actively under development

Active projects have seen activity in the project timeline (wiki, commits, tickets, forum, etc) within the past x days. Should a project not see activity within this timeframe, it falls into the 'dormant' category.

2) Dormant - under development (just not right now)

The project admins are notified that their project is about to be labelled as 'inactive'. After y days of timeline inactivity, a dormant project is labelled as inactive.

3) Inactive - not under development at this time

An inactive project is just that: it's labelled as 'inactive'.


Notes:

The inactive and dormant statuses can be reversed to 'active' again simply by virtue of project timeline activity.

As production projects are still useful beyond development, they could be excluded from 'inactive' status completely. Even though the labels are categorized along 'development activity', having a production project labeled as 'inactive' could send the wrong message.

I think the expiration times for both active and dormant can be made constant - something like 30 days each. This way, two months of inactivity rolls a project into the 'inactive' category, unless the project is touched. I think we could make a concession given the status of the project, like giving 'alpha' projects a tighter deadline, and 'beta' projects less so.

I personally feel that its important to avoid terminology that would label a project in a strongly negative way, like 'abandoned'. Having a whole slew of 'abandoned' projects could reflect very poorly on dsource.org otherwise.

The most important metric here, IMO, is trac timeline activity. I don't think the other metrics (aside from the project status) hold any real value as to the vitality of the project overall. Things like SLOC or size of a commit are just too qualitative for a rating such a thing. Smile
_________________
-- !Eric.t.Anderton at gmail
Back to top
View user's profile Send private message Yahoo Messenger
JoeCoder



Joined: 29 Oct 2005
Posts: 294

PostPosted: Sun Jun 04, 2006 12:50 pm    Post subject: Reply with quote

You might also take into consideration forum activity. For example if a project is undergoing a major planning phase with less source code being written.

However, what if a project only uses dsource for some of its needs? For example, a separate domain for better visibility to search engines (or they just need a server with custom software, yet use dsource for svn), or the project grows to such a size that several categories are needed on the forums? So maybe source code changes should be the primary metric.

Now, I'm still relatively new here, so feel free to overlook any of the areas where I don't know what I'm talking about. But I suppose that the more accurately you measure activity, the less meaningful your results will be, and vice versa.
Back to top
View user's profile Send private message
brad
Site Admin


Joined: 22 Feb 2004
Posts: 490
Location: Atlanta, GA USA

PostPosted: Sun Jun 04, 2006 3:45 pm    Post subject: Reply with quote

JoeCoder wrote:
But I suppose that the more accurately you measure activity, the less meaningful your results will be, and vice versa.

No lie there!!

BA
Back to top
View user's profile Send private message
clayasaurus



Joined: 21 May 2004
Posts: 857

PostPosted: Mon Jul 17, 2006 9:39 pm    Post subject: Reply with quote

I agree with pragma on the active, dormant, and inactive tags projects can be given. I think another tag that could be added is 'maintained.' The project authors should be allowed to choose which tag to give their projects.

I think project activity stats from there can be ranked with a formula that derives a percentage based upon

1) Maturity - Age of project - As the project ages and still remains active, it should rank higher than a younger active project.

2) SVN activity - (Number of commits) * (size of commit) + last commit month/year

3) Popularity - Number of forum posts not posted by the author himself

Variables that I think do not matter are...

- checkouts, downloads, page hits : Not reliable enough. Someone can post their project link to slashdot or a forum to make their project rank higher.

- # of wiki pages : not necessarily a factor in determining anything about a project

Also, I think the stats can be updated twice a year to help avoid 'stat tweaking' where someone tweaks thier project just to see their stats raise, twice a year is less dramatic
Back to top
View user's profile Send private message AIM Address
Display posts from previous:   
Post new topic   Reply to topic     Forum Index -> Site All times are GMT - 6 Hours
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB © 2001, 2005 phpBB Group