Wikipedia:Wikipedia Signpost/2011-03-07/Technology report

February Engineering Report published
(Due to a change in titling, this is in fact the second February update to be published. In future, all reports will cover the events in the month named in the title.) The Foundation's Engineering Report for February was published last week on the Wikimedia Techblog, giving a brief overview of all Foundation-sponsored technical operations in the last month. It summarised the developments:
 * the racking [i.e. setup] party at our new data centre in Virginia
 * the Data Summit that happened in early February in California
 * the release of Editor Trends study data and tooling
 * the painful, but ultimately successful, deployment of MediaWiki 1.17 to all Wikimedia wikis.

The update also noted that a job opening had been posted for a contractor in the Netherlands to support the Operations team in designing and maintaining the Wikimedia network(s), and perform on-site work in the data centre facilities in Haarlem and Amsterdam. The Foundation also noted its intention to hire a "Rich Text Editor Engineer", an indication that the Foundation is serious about its desire to provide its own WYSIWYG in-place functionality, a project for which research has begun (for context, see previous Signpost coverage). This might entail a move away from Wikimedia's traditional revise-and-save model to a more Google Wave-like approach, added developer Trevor Parscal. (On a related note, the report also discussed a new JavaScript parser for wikitext using parsing expression grammar.) In other news, Sumana Harihareswara was hired as a contractor to help out with Google Summer of Code 2011 and the Berlin Developer meeting.

In reference to the new Virginia data centre, the Foundation noted that all that was left was "finishing touches" to the hardware arrangement, as well as the initial setup of the software, "configuration of the first clusters of servers and services" and "network transport and transit services to be installed". In addition, contractor Russell Nelson has installed and deployed Swift on a test cluster of three machines. This forms part of the WMF's intent to improve the media storage architecture; the next steps are "fixing some bugs and doing some preliminary testing". The area of backups and general data redundancy has also seen significant developments: the operations team "have purchased a dedicated storage solution which will arrive in March... Once servers in the new data centre are online, and our private connection between Tampa and Ashburn is up, we will be able to replicate all data between the two data centres as well." Discussing the LiquidThreads project, the report also explained that "documentation on upcoming back-end and architecture changes [and] design specifications have been published".

The Foundation also announced the start of work done on two projects that have traditionally generated a great deal of debate: a system to allow users to censor their own visits to Wikimedia sites, and a mechanism for allowing expert reviews of articles. For the former, the report noted that initial UI design recommendations had been drawn up; on the latter, the report noted that a set of "draft requirements" had been drawn up for an "open review system for Wikipedia, as well as an API and user interface for quality indicators". The report, the Foundation's Engineering update to date, also noted work in a number of other areas not covered in this summary.

Brion Vibber rehired
The WMF's current Chief Technical Officer (CTO), Danese Cooper, has announced the rehiring of former CTO Brion Vibber to the post of Lead Architect. The post will be in the second layer of the current employee hierarchy, and Brion will start on March 31, 2011, she reported (Wikimedia Techblog).

Brion's name will be familiar to many Wikipedia regulars; indeed, in acronym form he gives it to this very report. The author of much of the original code in MediaWiki, and, as Wikimedia's first paid employee, having been among its most involved programmers for a number of years, Brion left the Foundation in 2009. He joined StatusNet, an open source startup focused on microblogging, while remaining active as a Wikimedia volunteer (see previous Signpost coverage). Danese explained Brion's new role:

Brion's first project will be on the team tasked with re-writing MediaWiki's parser, which should be both a challenging and rewarding effort, to which Brion tells me he's looking forward.

In a blog post, Vibber outlined this "next-generation parser work" briefly, saying that it will involve separating "weird template edge cases" from those that can be treated more easily.

In brief
Not all fixes may have gone live to WMF sites at the time of writing; some may not be scheduled to go live for many weeks.
 * An issue where the presence of a large SVG would break any page it was embedded in has now been fixed (bug #27508).
 * Two dates have been proposed for the 2011 Berlin developer Meeting in mid- or late May, May 20–22 and May 13–15. A straw poll is being run to decide which should be used.
 * Users on mobile devices will now be redirected much more quickly to prevent excessive loading times (bug #27690).
 * Developer Mark Hershberger has signalled his intention to ensure that one of last year's GSoC projects, "Reasonably efficient interwiki template transclusion" (see previous Signpost coverage), comes into production (wikitech-l mailing list).
 * Users of newer browsers will now be informed that their file exceeds the filesize limit (100MB on Wikimedia wikis) before the upload commences, rather than after it finishes (bug #26217).