Tuesday 17 February 2009

Wavemaker 32 bit rpm on amd64 debian

As there is no appropriate .deb file for wavemaker but the rpm is pure java so the following works:

rpm2cpio wavemaker- >wm.cpio  
cpio -idv  < wm.cpio

mv opt/wavemaker- /opt/   /opt/wavemaker-

Sunday 15 February 2009

£900m googlewhack

In Saturday's Guardian Jill Treanor treated us to an 'explanation' of the downfall of Lloyds, which left me perplexed and annoyed, so I turned to Google to try to decipher the jargon.

I discover that a "policy tax charge" is a googlewhack so am none the wiser.

Friday 13 February 2009

"Yahoo!", the number two, plays a spoiling game

I went to hear Mr Raghu Ramakrishnan of Yahoo! Research give a talk at ComLab, which was a real treat. Lovely people in a lovely building doing interesting things, this is the joy of living in Oxford and I am determined to take more advantage of it.

The talk was rather light on detail, if clearly and energetically given. The problem space was described very well, but the solution was mostly given by pointing to DBLife, so it can be done but the how is left for you to discover.

If, like me, you think of "Yahoo!" as a failing company with an incorrect belief that punctuation can form part of a name then you may not have paid much attention to them.
(Douglas Crockford's json excepted).

However the point was well made that as the number two in the market "Yahoo!" must, necessarily, play a market disrupting game and the way they are doing this is to open their API. The current monetised search model is sown up by big G. The next generation will be 'semantic' search:

- the search engine inferring the reason for the search and possibly tracking episodes in a search which may be engaged in over a number of sessions over a period of days or months.

- the search engine extracting more structure from the pages crawled.

How to do this is unknown so "Yahoo!" have opened their API, lowered the barrier to entry for academics and others to play with the results of spidering. No more having to setup the spiders, parsers or vast datastores previously required to start a websearch project.

"Yahoo!" do not want anything in return, all they want is the current incumbent de-throned, so that they can have another shot, possibly in partnership with the next big idea.

Thursday 5 February 2009

Bibliomania as a search provider

A nasty bug surfaced , which I introduced to poem in 2007, when Bibliomania was reunited with current melati/poem, see Melati Developers for gory details. The site is now even faster.

That done I have created an OpenSearchDescription for Bibliomania, which can be installed from the site or via the Mozilla addon repository.

I would like a few things to happen:

Someone with IE7 or IE8 try clicking on the little search icon at the bottom of the 'your details' login screen on Bibliomania.

Someone with Mozilla go to the Mozilla Addon site, search for 'Bibliomania search', create an account and login, install the addon and then review it positively. Ideally this will be more than one person, but at least one is required, other than me, before it is accepted.

You might also want to check the new entry about search in the Bibliomania FAQ.