This is getting stranger and stranger. Since my last entry the number of Google results for my tag string has dropped again, from 119 to 48.
Since updating my "tag string" on 2007-11-21 I have been monitoring the number of Google results for the new string. It gradually increased to about 150 (out of about 200 pages that include the string) and then mysteriously started decreasing till it was down to about 50. It was not that Google was dropping my pages entirely from its index (that would be less surprising) but it seems like it was reverting to earlier versions. That was how things stood a couple of days ago, but now it has suddenly shot up to 119 - I am confused.
Google really doesn't seem to like my website at all. It might be partly due to the fact that my pages are all XHTML but it is indexing them so that is probably not the whole story. Google is regularly downloading my valid sitemap file but is seems to be largely ignoring it. Until now I have not had a Google specific page on my website but I just created one.
I spent a little while analysing my Apache logs recently and noticed that my site had been accessed from a few interesting places such as:
There was also an access from GSi which is the UK Government Secure Intranet, a managed service operated by OGCbuying.solutions and delivered through a partnership with Cable & Wireless (formerly Energis Communications Ltd).
Until today, the title element of all these pages was "Ian Gregory" and a couple of people had pointed out that a page title should really be related to the subject of the page. Last night I modified my CMS so that I could set page titles to correspond to section headings ("About Me" in the case of this page). There were a few knock-on effects and I ended up making a number of changes - for the better I hope. One of the obvious changes is that there is now a single navigation page with two columns, one for the alphabetical list of topic links and the other for the chronological list of pages (these were previously on two separate pages). The navigation page also now uses its own style sheet.
I have switched to a new tag string (the 16 character random looking string in my page footers) and I will be monitoring how quickly Google picks up the change.
I have also added missing dates on some of my pages and now when you click on "Recent" in my page header, you get a list of pages ordered by date of "last entry" rather than last modified time on the source file. This means that pages do not move to the top of the list when I make minor edits in earlier entries. My XML sitemap still uses the last modified date though - so that search engines know that pages need fetching again.
About ten years ago, when I set up my first personal website the main purpose of it was as a place to store my bookmarks so that I could access them from anywhere with web access. Over the years the site grew and became much more than a list of bookmarks and this page is part of what the website eventually became. If you have been browsing around you will be familiar with the structure but basically the site is divided into sections by topic, and within each section there is at least an "Intro" page and a cronological "Notes" page. Until today there was also a "Links" page in each section, and taken together, all those links were effectively my online bookmarks.
Well, I have just spent a couple of days reviewing all those links and combining all the best ones into a single file in Netscape Bookmark File Format for import into my del.icio.us account. Once I had got them all in there and tagged them appropriately I then deleted all the "Links" pages on this site and removed the links to them. I used RedirectMatch in a .htacces file to redirect any requests for my old "Links" pages to my del.icio.us profile, which can also be accessed from my home page. That is enough for now, but eventually I will import another batch of bookmarks - ones that were not on my "Links" pages.
The problem I reported yesterday is (partly) solved. I posted a question in the Google Custom Search Help Group and received a prompt response from "omr". It is a known limitation which is documented here:
If your Custom Search Engine includes more than three sites, the results may be from a subset of our index and may differ from the results of a 'site:' search on Google.com.
I was including five sites. After reducing it to three I started getting expected results. I still have the problem that my pages are not being crawled nearly as frequently as I would like but there I expect I will just have to be patient.
For some reason my Google custom search facility is finding far fewer of my pages than a conventional Google search, but I don't have time to investigate right now. Even a conventional Google search is not finding all my pages, despite the fact that they are all listed in up to date sitemaps that Google is downloading quite regularly, and looking at the pages that are indexed by Google, many are a couple of months out of date. This is annoying but there is probably not much I can do about it.
I have made a few changes to my website recently. One thing is that I have changed the DOCTYPE declaration on all these pages from XHTML Mobile 1.0 to XHTML Basic 1.1 - a candidate recommendation of the W3C which is intended to be the convergance of XHTML Basic 1.0 and the Mobile Profile.
Another change is in the top navigation bar. I have renamed the first two links from "Home" and "Page List" to "Contents" and "Recent". The third link used to be "Contact" and went to my contact form which was in the same section of my XHTML site as this page. I have changed that to "Home" which actually links back out of the XHTML site to my HTML entry page. I edited the contact form itself, changing it to HTML and moved that as well. This is all a concession to people using Microsoft Internet Explorer which is still incapable of viewing XHTML.
I have gradually started putting up more content outside of the zenatode site. For example I have been blogging a bit on my Multiply site, using Digg and YouTube, bookmarking stuff at the Internet Archive and have even (reluctantly) created a MySpace page. All this stuff is now linked from my HTML Home Page.
At one point I cobbled together a basic search facility for this site but when I made a major change to my homebrew CMS I removed the it. I intended to create an improved local search facility but I never got round to it, thinking that people could always do a Google site search if they were really interested in searching my site. Well now I have added a Google Co-op account to my basic Google account and created a Custom Search Engine called zenatode search which searches not only my zenatode site, but also a few closely related ones.
Just to follow up on my previous entry, my banana photo has still been getting a lot of hits (over 2000 since October 18th). The top eight referring domains are (with number of hits in parentheses):
I just used Comic Life (which came bundled with my new iMac) to create a doctored version of the photo which I have substituted for the original!
In the last six months my banana photo has been downloaded over 7,000 times, largely as a result of inline linking from various web forums and myspace profiles. The photo was even used as the basis for a caption competition on the bulletin board of the Royal Holloway Students' Union and some of the suggested captions are pretty funny.
One more new feature of my website is that it is possible to address individual dated entries, like the one on this page about my 2004 climbing trip to the Alps. All date headings associated with dated entries are now identified with an id attribute of the form dyyyy-mm-dd. The the leading "d" (for date) was necessary because the value of an id attribute is not permitted to start with a digit.
Yesterday I noticed that my contact form was not working and further investigation revealed that it would not have been working since 2006-06-14 when Black Cat Networks upgraded Exim and Apache on their hosting servers. It was easy enough to fix and it is working again today.
I am no longer working at the University of Hertfordshire and have been updating my website where necessary to reflect my new situation.
I noticed that I am listed under Ian Gregory on the Tell Me About UK Directory. Their list is not as good as mine though!
As well as continuing to add stuff to existing notes pages I have also done a bit of reorganisation and added four new main sub-sections: Linux, Science, Religion and The Arts. In my Internet sub-section I have added pages about email and broadband, and under Global Issues there is a new Palestine page.
Although my site is (hopefully) fairly easy to navigate, it is now big enough that you might struggle to find something by casual browsing. Fear not, there is a search facility on the zenatode homepage.
A few days ago I Googled for "Ian Gregory" and started looking through the 4,400 results. It was clear that a large number of these entries referred to a fairly small number of individuals and I ended up creating this page to summarise what I found.
In a related move, I submitted my website to the dmoz open directory project, under the Personal Homepages category. I do not expect the submission to be rejected but it might take some time to be approved.
I have done some fairly extensive tinkering with my website recently. There are a couple of new directories, one for stuff about Aidan Andrew Dun which was previously in the "Miscellaneous" directory and one about UK peace activism which incorporates some stuff from the old "Issues" directory. The "Issues" directory itself has been renamed "Global Issues" to reflect a new focus, but the URL remains the same to minimise link breakage. Sorry for any disruption, but as they say, you can't make an omelette without breaking eggs.
I have added a new directory for local events and issues, changed the title of some directories (URL's are not affected so links to the site will not get broken) and spruced up my resume.
qtrbksivbbadwdab www.zenatode.org.uk/ian/ Ian Gregory 2008
