Hello All,

On Wednesday 3rd February, a hard disk failure occurred on one of the
two web servers used to host the Quranic Arabic Corpus. This has lead
to a reduced service on the website over the last 24 hours. The entire
website is working correctly, with the exception of not being able to
post new messages on the message board. This last issue is under
investigation, and I hope to help to get this resolved soon. However,
viewing existing messages, and making use of the word-by-word
grammatical information is also working, as is the rest of the
website. My apologies for this, hopefully this will be up and running
again shortly. Despite this technical issue, the website has
successfully served 11,637 user page views with no down-time during
the last 48 hours.

I think that is important to point out that the server and database
hosting required to power the dynamic website is all run quite
professionally, but on a volunteer basis as part of a sub-domain under
quran.com. The kind efforts and technical expertise of the team behind
quran.com have meant that the corpus website was up and running again
quite quickly, and because two web servers are used for backup
redundancy, there was no loss of service at any time. My kind and
sincere thanks go out to the team who run and maintain these web
servers, free of charge as a volunteer educational and humanitarian
effort to promote deeper understanding of the Quran.

The Quranic Arabic Corpus website is resource intensive, since much of
the grammatical information and syntactic images are generated
on-demand when a specific verse of the Quran is requested by a user
for the first time. The website is hosted on two dedicated web
servers, with the linguistic database hosted on a separate third
server. The hard disk failure has meant that messages for a single day
have also been lost (a few dozen grammatical corrections). Going
forward, this should not occur again, since more frequent twice-daily
backups of the message board will hopefully be put in place.

Most of the website's pages are generated dynamically by the two
hosted web servers from a set of core tags and features encoded in the
linguistic database. To give some idea of the scale of the website’s
dynamic information, if only the brief grammatical information was
printed, together with the morphological word segment images, at a
very small half font-size compared to the online web pages, then the
word-by-word interlinear format alone
(http://corpus.quran.com/wordbyword.jsp) would take up 5,530 A4 pages
- assuming no extra website banners or menus. This is equivalent to
about 10 large and dense volumes of printed grammatical information,
with detailed colored and segmented images for each of the Arabic
words in the Quran. These 10 large printed volumes would exclude the
treebank and many other parts of the website.

I hope that volunteer annotators would please be patient for a short
while, until full message board functionally is restored. Please also
accept my apologies if you have to register and sign-in through the
website again, hopefully this should not happen again. I will be sure
to let you all know when things are fully up and running. Once things
are working, it would be very kind if corpus annotators might consider
reposting the last few dozen or lost corrections to words.

Kind Regards,

-- Kais Dukes

Language Research Group
School of Computing
University of Leeds

http://corpus.quran.com - The Quranic Arabic Corpus
comp-quran@comp.leeds.ac.uk - Computational Quranic Arabic discussion list

Reply via email to