It is precisely in knowing its limits that Philosophy existsImmanuel Kant
Also relates to XUL
This entry is a little late in the day, since with massive work commitments I forgot to post it, but on 14th April, Amazon released a beta version of its new search portal, A9.com. As well as the search results (supplied by Google), the portal offers two additional panels - Book Results shows results from an Amazon search on the keywords; Search History provides a chronological and actual history of keywords searched. The latter feature can be integrated with the browsing history offered by the A9 Tool Bar (IE only!) to offer quite a powerful tracking mechanism for intensive searches. A9 also offers a simple query mechanism direct through the URL by simply typing the keywords after the base URL (eg <a href="http://a9.com/severn solutions">http://a9.com/severn solutions</a>) - beneficial for those browsers that don't offer Quick Search features.
John Battelle discusses the Amazon/Google relationship in this venture, and the potential opportunities provided by the search history facility, while Resource Shelf plucks out some of the current weaknesses of the current beta release.
The Tool Bar offers a number of additional features. However, as is so often the case with search tool bars, it is only compatible with IE. Perhaps, someone has already followed the step taken by relatively new Eurekster to produce a XUL version of the tool bar for the blossoming number of Mozilla and co users. As for the design of the portal interface with three resizable panels: It is clear and usable on the major web browsers, but perhaps accessibility could be improved - Lynx illustrates the reliance of the panels on client-side scripting where they fail to work.
Posted on Apr 24, 2004 at 14:07:18. [Comments for New Amazon Search Portal- 0]
This post does not relate to any other topics
Just picked up on a couple of new features added to Google in its attempt to maintain dominance in a rapidly evolving search engine market.
With talk of Longhorn integrating personal data collection for use by the MSN search engine, Yahoo attempting to provide tailored information, and the emergence of new engines like Eurekster with a major emphasis on personalized results and social networks, a more personalised experience seems to be a major objective for the big players now.
Posted on Apr 04, 2004 at 19:39:04. [Comments for Latest Google Features- 0]
Also relates to Apache
With Windows 98 as my primary OS on my current system, I have never really delved deep into Apache, except a previous failed attempt to establish Virtual Named Hosts. Well, today I returned to this with renewed vigour on reading Simon Willison's post and accompanying links on mod_rewrite. This voodoo is such a powerful feature for improved URI-security, search engine optimisation and usability. However, while I was able to play around with a few trial regular expressions on my test platform, conflict soon arose with the use of relative URIs in test builds to accomodate their directory location within subfolders of the root. So I decided it was time to finally get Name-based Virtual Hosts.
This was no easy feat - understandably there is little documentation online for Windows 98 and Apache! - and has taken a few hours to get to grips with. However while the solution may not be correct, it seems to work well now, and I can directly clone work-in-progress sites to the finished article. It also means I can take advantage of mod_rewrite safe in the knowledge that what I create on the test platform can be mirrored to the public server. Here are a few links that came in handy along the way:
The Windows 98 process boils down to:
127.0.0.1 newhost
The first entry will always be the same and the second the name of the new host. I found spaces were required between each field on the line rather than tabs.
<VirtualHost *>
ServerName newhost
ServerAdmin webmaster@newhost
DocumentRoot "C:/Apache/htdocs/newhost"
</VirtualHost>
Of course additional directives could be added to the Virtual Host, but in this case I am just accepting the defaults defined for the main server (ie localhost) in Section 2 of the config file.
I also found I needed to disable the Internet Connection in Internet Options of Internet Exploder to prevent the Connect window form popping up (annoying as much as anything). So by having two HOSTS files, one with the domain I am working on, www.tomwright.me.uk, pointed at the localhost, and one excluding the domain, I can work in the development environment on my test platform and transfer this as is to the public server without any issues relating to URI paths. And, of course, finally get to really play with mod_rewrite. A quick example being the following code in the .htaccess file at the root of a particular Virtual Host:
RewriteEngine On
RewriteRule solution/([a-z]+) /services.php?service=$1
Which cloaks the GET request for each service presenting a more usable and Search Engine friendly URI presented as a directory structure.
Perhaps one day I might actually trade in my Win98 and take the plunge into RHL….
Posted on Sep 03, 2003 at 01:39:23. [Comments for Virtual Hosts In Win98- 0]
Also relates to Accessibility
Scribbling.net offers an interesting article on optimising content for Googlebot (the Google Search Engine indexing software). It demonstrates the similar thinking that goes into SEO and building for accessibility. Search engine software is really the ultimate disabled user, limited by it's specific algorithms.
The point regarding the relationship between an anchor and anchor text is of obvious significance here. I had already become aware this could create problems when I first ran these pages through Bobby. It automatically picks up on repetitive anchor content (eg the use of permalink). I can see the point - a visually impaired user listening to their speech browser could get quite confused with the same anchor name repeated over and over. Noted for change.
Posted on Jul 03, 2003 at 23:30:14. [Comments for Accessibility For Google- 0]