RE: Project Activity
Sometimes lists are where the activity is, commits alone don't credit the essential design and planning effort put in by users commiters and non-commiters that shapes the product and maps its progress. d. This makes me think of all the projects on SourceForge that shoot up high into the project ratings, with a high activity percentile, just because they turn the news function into a message board. -- To unsubscribe, e-mail: mailto:[EMAIL PROTECTED] For additional commands, e-mail: mailto:[EMAIL PROTECTED]
Re: Project Activity
On Mon, 13 May 2002 20:17, Danny Angus wrote: Sometimes lists are where the activity is, commits alone don't credit the essential design and planning effort put in by users commiters and non-commiters that shapes the product and maps its progress. Agreed - even worse. Sometimes after these activity meters turn up you get committers breaking up one commit into many commits, presumably to push their activity level up. You also get the many typographic changes for much the same reason. I have found that higher healthy activity is actually indicated by small localized changes. This is not going to be captured in a simple count the commits and note the committer style approach. -- Cheers, Peter Donald -- To unsubscribe, e-mail: mailto:[EMAIL PROTECTED] For additional commands, e-mail: mailto:[EMAIL PROTECTED]
RE: Project Activity
I don't really think project activity can be calculated from code changes alone. I would say project activity should cover the complete process of setting up requirements for the software, designing, implementing, testing and using it. If Jakarta would have a structured software development process this would be easy to do, but because of the all the diverse ways of coming to a release this is hard to do I guess. But maybe some milestones/checks can be put up which could be measured (e.g. in respect with time between the milestones, time between a bugreport and a bugfix). This way you might create a couple of vague notions like: *time-to-release (short/medium/long) *stableness (amount of bugs reported, hihg/medium/low) *userbase (large/medium/small) *amount of minor releases (bugfix release) per month or year. Those kind of stats might give a user way more information than the amount of commits or changes to a certain file. I wouldn't even want to know ;-). Project activity measurements only create unnecessary competion IMHO. In Jakarta's case projects are rejected anyway if they don't have a certain activity. Alef -Original Message- From: Peter Donald [mailto:[EMAIL PROTECTED]] Sent: Monday, 13 May 2002 12:31 To: Jakarta General List Subject: Re: Project Activity On Mon, 13 May 2002 20:17, Danny Angus wrote: Sometimes lists are where the activity is, commits alone don't credit the essential design and planning effort put in by users commiters and non-commiters that shapes the product and maps its progress. Agreed - even worse. Sometimes after these activity meters turn up you get committers breaking up one commit into many commits, presumably to push their activity level up. You also get the many typographic changes for much the same reason. I have found that higher healthy activity is actually indicated by small localized changes. This is not going to be captured in a simple count the commits and note the committer style approach. -- Cheers, Peter Donald -- To unsubscribe, e-mail: mailto:[EMAIL PROTECTED] For additional commands, e-mail: mailto:[EMAIL PROTECTED] -- To unsubscribe, e-mail: mailto:[EMAIL PROTECTED] For additional commands, e-mail: mailto:[EMAIL PROTECTED]
Re: Project Activity
Agreed! Danny Angus wrote: Sometimes lists are where the activity is, commits alone don't credit the essential design and planning effort put in by users commiters and non-commiters that shapes the product and maps its progress. d. This makes me think of all the projects on SourceForge that shoot up high into the project ratings, with a high activity percentile, just because they turn the news function into a message board. -- To unsubscribe, e-mail: mailto:[EMAIL PROTECTED] For additional commands, e-mail: mailto:[EMAIL PROTECTED] -- To unsubscribe, e-mail: mailto:[EMAIL PROTECTED] For additional commands, e-mail: mailto:[EMAIL PROTECTED]
Re: Project Activity
On Mon, 13 May 2002, Peter Donald wrote: On Mon, 13 May 2002 20:17, Danny Angus wrote: Sometimes lists are where the activity is, commits alone don't credit the essential design and planning effort put in by users commiters and non-commiters that shapes the product and maps its progress. Agreed - even worse. Sometimes after these activity meters turn up you get committers breaking up one commit into many commits, presumably to push their activity level up. You also get the many typographic changes for much the same reason. Breaking one big commit into many commits is not bad. It makes things easier to review, the commit comment can describe much better what has been done in the file. Putting a 'ranking' on commiter's activity is however very bad. Some are working full time ( as part of their job ), some are using the little free time they find ( or sleep less ). I think the second category deserves a lot of apreciation, even if they may have fewer commits. Costin I have found that higher healthy activity is actually indicated by small localized changes. This is not going to be captured in a simple count the commits and note the committer style approach. -- To unsubscribe, e-mail: mailto:[EMAIL PROTECTED] For additional commands, e-mail: mailto:[EMAIL PROTECTED]
Re: Project Activity
Many commits without description and many cosmetic changes. Hey now, my commits are coming through with no message because of a bug in NetBeans. It's got nothing to do with the stats. Some of those 'cosmetic' changes like checkstyle issues are LONG overdue. -- dIon Gillard, Multitask Consulting Work: http://www.multitask.com.au Developers: http://adslgateway.multitask.com.au/developers Nicola Ken Barozzi [EMAIL PROTECTED] 05/13/02 10:00 PM Please respond to Jakarta General List To: Jakarta General List [EMAIL PROTECTED] cc: Subject:Re: Project Activity From: Peter Donald [EMAIL PROTECTED] On Mon, 13 May 2002 20:17, Danny Angus wrote: Sometimes lists are where the activity is, commits alone don't credit the essential design and planning effort put in by users commiters and non-commiters that shapes the product and maps its progress. Agreed - even worse. Sometimes after these activity meters turn up you get committers breaking up one commit into many commits, presumably to push their activity level up. You also get the many typographic changes for much the same reason. This is exactly what has happened to turbine-maven just after the statistics were made. Many commits without description and many cosmetic changes. Measuring how well a project is doing with these stats is nonsense. There is no semantics in numbers. Say you are having tons of letters from angry users that claim that your product sucks. Is the number of posts still a health indicator? Maybe of the mailing list software ;-) -- Nicola Ken Barozzi [EMAIL PROTECTED] - verba volant, scripta manent - (discussions get forgotten, just code remains) - -- To unsubscribe, e-mail: mailto:[EMAIL PROTECTED] For additional commands, e-mail: mailto:[EMAIL PROTECTED] -- To unsubscribe, e-mail: mailto:[EMAIL PROTECTED] For additional commands, e-mail: mailto:[EMAIL PROTECTED]
[EGO] Re: Project Activity
FWIW, there is no 'ranking' in the cvs activity report, unless you consider alphabetical order ranking. But I'm not going to complain about people who would otherwise do nothing doing some of the 'cosmetic' stuff like documentation/javadoc etc. And everyone seems to have ignored the file activity report - which helps to find code that is unstable. -- dIon Gillard, Multitask Consulting Work: http://www.multitask.com.au Developers: http://adslgateway.multitask.com.au/developers [EMAIL PROTECTED] 05/14/02 01:33 AM Please respond to Jakarta General List To: Jakarta General List [EMAIL PROTECTED] cc: Subject:Re: Project Activity On Mon, 13 May 2002, Peter Donald wrote: On Mon, 13 May 2002 20:17, Danny Angus wrote: Sometimes lists are where the activity is, commits alone don't credit the essential design and planning effort put in by users commiters and non-commiters that shapes the product and maps its progress. Agreed - even worse. Sometimes after these activity meters turn up you get committers breaking up one commit into many commits, presumably to push their activity level up. You also get the many typographic changes for much the same reason. Breaking one big commit into many commits is not bad. It makes things easier to review, the commit comment can describe much better what has been done in the file. Putting a 'ranking' on commiter's activity is however very bad. Some are working full time ( as part of their job ), some are using the little free time they find ( or sleep less ). I think the second category deserves a lot of apreciation, even if they may have fewer commits. Costin I have found that higher healthy activity is actually indicated by small localized changes. This is not going to be captured in a simple count the commits and note the committer style approach. -- To unsubscribe, e-mail: mailto:[EMAIL PROTECTED] For additional commands, e-mail: mailto:[EMAIL PROTECTED] -- To unsubscribe, e-mail: mailto:[EMAIL PROTECTED] For additional commands, e-mail: mailto:[EMAIL PROTECTED]
Re: Project Activity
I think this might be what you mean by file activity, but might not. Aren't there two types of project activity. Developer activity and user activity. Having no developer activity is not a bad sign if there is high user activity, ie) sign of a solid/stable piece of code. Now I don't know of any like that, but maybe it's just lack of looking :) On Sun, 12 May 2002 [EMAIL PROTECTED] wrote: Back in mid March there was a discussion around the jakarta overview ( http://jakarta.apache.org/site/overview.html ) and what / wasn't a good measure of project activity. My comment back then was commits on a project are a good indicator. So anyway, I've added this reporting into the cvs head of maven. For a sample, see: http://jakarta.apache.org/turbine/maven/maven-reports.html In particular the change log, developer activity and file activity. -- dIon Gillard, Multitask Consulting Work: http://www.multitask.com.au Developers: http://adslgateway.multitask.com.au/developers -- To unsubscribe, e-mail: mailto:[EMAIL PROTECTED] For additional commands, e-mail: mailto:[EMAIL PROTECTED] -- To unsubscribe, e-mail: mailto:[EMAIL PROTECTED] For additional commands, e-mail: mailto:[EMAIL PROTECTED]
Re: Project Activity
I think that's pretty cool functionality. I don't agree with your way of judging project activity.. Thats like trying to put a metric on how blue the sky is today. I'm sure you could but what of the day where your metric says its grey but I'm looking up and seeing blue (perhaps because a grey day in June would look blue in January). Its all subjective. I don't believe in software engineering anymore. -Andy On Sat, 2002-05-11 at 11:07, [EMAIL PROTECTED] wrote: Back in mid March there was a discussion around the jakarta overview ( http://jakarta.apache.org/site/overview.html ) and what / wasn't a good measure of project activity. My comment back then was commits on a project are a good indicator. So anyway, I've added this reporting into the cvs head of maven. For a sample, see: http://jakarta.apache.org/turbine/maven/maven-reports.html In particular the change log, developer activity and file activity. -- dIon Gillard, Multitask Consulting Work: http://www.multitask.com.au Developers: http://adslgateway.multitask.com.au/developers -- To unsubscribe, e-mail: mailto:[EMAIL PROTECTED] For additional commands, e-mail: mailto:[EMAIL PROTECTED] -- http://www.superlinksoftware.com - software solutions for business http://jakarta.apache.org/poi - Excel/Word/OLE 2 Compound Document in Java http://krysalis.sourceforge.net/centipede - the best build/project structure a guy/gal could have! - Make Ant simple on complex Projects! The avalanche has already started. It is too late for the pebbles to vote. -Ambassador Kosh -- To unsubscribe, e-mail: mailto:[EMAIL PROTECTED] For additional commands, e-mail: mailto:[EMAIL PROTECTED]
Re: Project Activity
This makes me think of all the projects on SourceForge that shoot up high into the project ratings, with a high activity percentile, just because they turn the news function into a message board. If I were to move code from src/java to src/foo, that's quite a few commits. You written your 1000 lines today? :D - Leo On Sat, 2002-05-11 at 17:07, [EMAIL PROTECTED] wrote: Back in mid March there was a discussion around the jakarta overview ( http://jakarta.apache.org/site/overview.html ) and what / wasn't a good measure of project activity. My comment back then was commits on a project are a good indicator. So anyway, I've added this reporting into the cvs head of maven. For a sample, see: http://jakarta.apache.org/turbine/maven/maven-reports.html In particular the change log, developer activity and file activity. -- dIon Gillard, Multitask Consulting Work: http://www.multitask.com.au Developers: http://adslgateway.multitask.com.au/developers -- To unsubscribe, e-mail: mailto:[EMAIL PROTECTED] For additional commands, e-mail: mailto:[EMAIL PROTECTED] -- To unsubscribe, e-mail: mailto:[EMAIL PROTECTED] For additional commands, e-mail: mailto:[EMAIL PROTECTED]
