[Bioc-devel] bug tracker (was: graph package conflict with rtracklayer)

[Cook, Malcolm]

Hi Martin (and others with opinions on bug/issue tracking for BioConductor projects)

> > Also, off topic, but, is there any advice on tracking status of
> > issues with bioconductor packages other than emailing to the list and
> > package authors?  The CPAN (for perl) offers an 'rt bug tracker'
> > which forwards posted issues to developers.  Something like that
> > might be a grand addition!
> 
> Hi Malcolm -- tried to disentangle the bug tracker from the other parts
> of the post.
> 
> There is no overall bug tracker; it is worth exploring, though no
> promises. Any feedback from other developers?

None.  Perhaps due to my inquiry being off-topic.  Perhaps your dis-entangling will bring some opinions out.  Thanks for that!

>  I think my reservation is
> that only some would adopt the practice, meaning that the developer and
> the conscientious bug reporter interested in knowing whether an issued
> had been reported / resolved would have to look both on the list and in
> the tracker.

The payback from such an effort could potentially 
	shield developers from too many duplicate reports
	shield users from annoying talented developers with bug reports that have already been reported and for which a workaround might exist
	provide tools from which library 'release notes' might be crafted/extracted (i.e. "in current release, the following bug were fixed (issues were addressed/features were implemented):"
	allow integration with source code control, issue/bug management, and importantly, wiki (i.e. 'trac' supplies some of this http://trac.edgewall.org/wiki/VersionControlSystem)

I share your reservation, but, honestly, would be willing to be conscientious and lead by example in regard to using it.  Of course not all developers would want to use such a beast, and those who do might use it to different degrees and with different effect.  In my opinion, if only three of the key packages I'm interested in tracking had issues logged  systematically, I would think the effort well worth it.  

Of course my tune might change once I become a full-fledged submitter to BioC.  Time will tell.  

> Thanks for the suggestion! Martin

Very welcome.  

I was also one of the people who pushed for RSS feed coming from commits to svn, and have made suggestions (https://stat.ethz.ch/pipermail/bioc-devel/2011-May/002624.html) on how to do it .  I am very thankful for the feed and read commits avidly, (esp to RNASeqGenie, IRanges, AnnotationDbi, Rsamtools, ShortRead, BSgenome, BioStrings).  So, I am biased toward facilitating tracking changes to project artefacts, such as bugs, issues.  IMHO, It would be GREAT if the commit log had (optional!) hyperlinks to the bug/issue they resolve/address.    But, it is bunch of infrastructure to cobble and if no one will use it, then, Oy, what a waste.  But if we all _were_ conscientious, both developers and user/analysts, they Oy, the joy.

My two cents,

Malcolm