[Building Sakai] [sakai2-tcc] Assignments 2

[May, Megan Marie]

No one suggested doing a drop in replacement so let's try to get past this.

This all started as a discussion on whether or not the community should pursue adding it to core.   As I said earlier, I am not married to one solution over the other.  This community *does* need to make a decision and that should be accomplished by analysis on what work would be involved should we go one way or the other.

Once a decision is made we can address the other concerns.

From: sakai-dev-bounces at collab.sakaiproject.org [mailto:sakai-dev-bounces at collab.sakaiproject.org] On Behalf Of Charles Severance
Sent: Friday, March 08, 2013 11:47 AM
To: sakai-dev
Subject: Re: [Building Sakai] [sakai2-tcc] Assignments 2


On Mar 7, 2013, at 4:01 AM, Adrian Fish wrote:


The Profile2 precedent that Megan offered is a little bit off the mark. Functionally, A2 does not offer enough above A1 to justify it being a replacement. Profile2 in functionally vastly superior to the Profile tool and was a no brainer replacement; you can't compare the Profile and Assignments1 scenarios.

The profile2 experience as evidence that A2 is a good idea is *way* off the mark.  On the contrary...

Steve replaced a very low functionality tool with a high functionality tool.   He built extensive transition apparatus.  He built forward and backward compatibility.   He provided tech support for every issue that was raised by the community and did so very quickly.   He dropped his own agenda when there was a problem with profile2.   He accepted feature requests that had little to do with his own local use cases, provided web services. etc etc etc.  He worked diligently for over a year making sure that P2 fit into Sakai.   Steve ran every version of P2 in his own production environment under load to be guinea pig for the rest of the community during development....  He released early and often and produced amazing documentation. It is an example of how things *should be* done.

In short, he was 110% committed to doing *whatever it took* to get from P1 to P2 in the eyes of the community.

We only accepted P2 in trunk when it was a real replacement and schools validated his work.

This is *very different from A2*.  A2 is not a complete replacement and there is no commitment to make it a complete replacement and no one willing to put their school / personal energy behind it doing whatever the community asks.  There is no commitment for a backwards/forwards compatibility.

All we have is the vague notion that if we drop it in trunk and mess up our code base then people will be forced to work on it - a horrible way to develop software.

The P2 experience *proves* that A2 is *not* ready and not likely to *ever* be ready for trunk unless there is a dramatic change in its current trajectory.  I am not criticizing A2 - it is likely a fine piece of software and makes its users very happy - it is just not ready for a drive-by trunk-commit.

We have *got* to stop talking about "X is a good idea" and get to "Who will go on record and behind X as along as it takes until the community is 100% satisfied."    If no one will do the work - we should drop the conversation about A2.  And just saying "if you put it in trunk, we promise we will do the work" is not enough either - we have seen that before.  If folks are wiling to make A2 a complete replacement - then do it in a branch - run it in your production environments - test the conversions / transitions / submit it to the community for review and fix the things the community identifies as issues.   Sound like a lot of work?  Yes - it is.   Dropping it into trunk does not lessen the work - it just makes it easier for people to not do the work - because it is already in trunk.

We have very bad experience in dropping software that is a partial solution in trunk hoping it will rally resources around the software.

/Chuck

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://collab.sakaiproject.org/pipermail/sakai-dev/attachments/20130311/8d357ef3/attachment.html