Saturday, 10 January 2009

The Configure Extension and Its Relevance

Recently I have started testing the Configure extension, it is very valuable to our project for a variety of reasons - and indeed quite unlike most of the other extensions we test.

Primarily it will help to ensure that testing is more comprehensive, if extensions are supported by configure then it reduces the chance that a software tester (such as myself) will miss out testing certain configuration settings.  Making sure that all configuration settings are tested at least independently, and even better together (of course, this isn't always feasible), is integral to the validity of results. Working by both reading the settings in the extension's documentation, and using Configure, ensures that this is done to the highest standard possible, given the software we have available.

Configure is also very useful to the end user, and this is one of the reason that I've included wether it is supported or not within the compatibility charts that I generate for tested extensions.  By knowing that it is supported the person who will be managing the wiki will know that the extension will be easier to work with and generally makes MediaWiki more accessible - one of the primary goals we have.

I certainly hope to see more extensions become supported by Configure, and a great deal already are, which is fantastic.  Developing support is also a fairly trivial process and does not consume a significant amount of time - for most extensions it will simply be the case of committing an array to the Configure extension settings file.

3 comments:

Mike.lifeguard said...

I think it's probably more significant that this extension represents a major change in the workflow for an entire /class/ of shell requests. Changes to configuration settings are perhaps the single most problematic area of server-side maintenance for end-users. Currently, these requests are fulfilled on an ad-hoc basis, nowhere near fast enough for the users' liking, and not necessarily in any particular order.

By shifting these tasks from the (primarily volunteer) shell users to stewards and/or bureaucrats, we will significantly increase the number of people fulfilling these requests, which will decrease turnaround time. As well, this will free up sysadmin time which might be better spent on other tasks - what those might be I'm not sure.

By focusing on the underlying issue and rethinking the workflow, we solve a major issue. An exercise one might test on other areas.

darkoneko said...

wait wait wait...

Does that mean that if aplied to the wikimedia wikis, stewards will be ablue to change some configuration setting ?

that seems scary, yet very awesome. I mean, not all steward have technological knownledge (I trust those of us who don't not to touch anything :) but that sure would speed up the simple requests.

kwj2772 said...

Special:Extensions doesn't work with Internet Explore. Until JavaScript errors is fixed, this extension shouldn't be installed on Wikimedia wiki.

And some extensions need SQL query, can be configured by system administrators.