Showing posts with label Open Directory Project. Show all posts
Showing posts with label Open Directory Project. Show all posts

Friday, 4 July 2008

Google PageRank for bloggers...

Note: this article has now moved to google-pagerank-for-bloggers on Cipher Mysteries

Over the past few months, I've been tracking where this Voynich News blog appears in Google's search results for "Voynich": it has been as high as #1 and as low as #13, which is quite a variation. It seems to lurch up to #5 (behind Wikipedia and voynich.nu, with two links each), before slowly sinking down to about #12 over the following fortnight, before cycling back up to #5 (or occasionally #1) again.

The other sites in the top 12 typically date from 2004 or before (equivalent to about 30 Internet years) and refer to each other a fair bit: and their age, interconnectedness, and accumulated inward links all work together to give them a Google PageRank of 5 or more when viewed as a network.

By way of comparison, Voynich News is part of a small new network of Voynich-related sites with few inbound links, and so has a PageRank of only 3. Furthermore, because the older sites are rarely (if ever) updated, their PageRank won't "flow" to Voynich News, whereas whenever Voynich News links to one of them, part of its PageRank flows out to supports theirs.

So why does Voynich News cycle around the first 12 places? I strongly suspect that to keep the PageRank values "fresh", every once in a while Google completely discards its PageRank cache and incrementally rebuilds it as it crawls around the Net. And so, without the impact of PageRank on its position, Voynich News would consistently be at #1-#5 (i.e. based solely on its content relevance). But as the PageRank cache gets progressively rebuilt, down it slides. :-(

To get a higher PageRank, I'd need people to mention it on their blogs and sites: but this rarely happens (Google also tries to filter out backlinks in comments, for example, as this was how an earlier generation of trolls tried to artificially inflate their PageRank). However, there's a big wave of Voynich interest coming later this year and forwards, so my only real crime is to be ahead of the game. Shame on me for being too early! ;-)

I've been updating this blog with (I hope) relevant and interesting Voynich-related stuff pretty much every day for ages now: but I've pretty much hit the limit of what I can do with Google and Blogger alone. For example, Blogger sites get penalised by Google because Blogger does not place different keywords or descriptions in the META tags in the header: but Google owns Blogger, for goodness' sake!

If Blogger simply added a set of META "keywords" based on the tags in a post, and had a configurable META "description" that only got put on the root page of the blog, I think that would be a huge improvement as far as SEO ("Search Engine Optimisation") goes. But Blogger seems more interesting in adding in useless sidebar Gadgets that nobody really benefits from, which is a shame.

I'd love to go in and turn Blogger round, as I think right now it's a wasted opportunity for Google: but it seems easier from where I'm sitting just to start my blog all over with a different blogging host, one which gets the SEO basics right. Which would be a shame, but there you go.

There's one little-known curiosity about META tags: if you don't have them on a web page (like most of the older sites in the Voynich network I mentioned above), then Google may well use the description given for the site in the Open Directory Project (though you can disable this with "NOODP" in a META tag). Seeing all the descriptions from the ODP I had recently edited suddenly popping up in the search listings for Voynich was quite a surreal experience for me (even though they have since disappeared again!)

My top blogger tips for high Google search results are therefore:-
  • Find out the PageRank of sites that rank highly on searches for your keyword - this is the value your blog will have to match in order to appear near the top most of the time (for comparable content, that is). You can find this out with the PageRank searcher from SEO Chat, or the SEOQuake plugin (for IE and Firefox).
  • PageRank is based on unique contentful inbound links to your blog (a repeated link in a site template, such as someone's sidebar list of blogs, makes little or no difference). How do you get these? Be nice to other bloggers lots of times, and hope some link back to you!
  • Getting from PR3 to PR4 (and then to PR5) will take time and effort. Even high quality inbound links from directories (such as the Open Directory Project, which people always used to think was helpful) may also not help because these often get duplicated around the Web - and when Google notices duplicates, it can reject all of them (pretty much).
  • Don't use Blogger, find a blog host who gives you proper control over META tags (sorry, Google/Blogger, but that's the way it is right now).

Saturday, 21 June 2008

The Open Directory Project...

I've just been accepted as an editor of the Open Directory Project: for several years, its Voynich Manuscript category has been fairly moribund (if not actually dead), and I thought it would be a positive thing to try to update and restructure it, so that it could provide a set of links to Voynich-related things that was actually useful - image galleries, research, theories, etc.

I made a good start at doing this: but then I ran into a heap of problems...

Firstly, it turns out that nearly every interesting Voynich-related site I've looked at over the past six years has stopped working: Voynich dead links significantly outnumber the live links. Furthermore, many dead links (even recent ones such as ms408.com) don't appear to have made it into the Wayback Machine. And "deeplinking" (linking to sites inside a website) into the Wayback Machine is a no-no in the ODP.

Secondly, while the intention of the Open Directory Project is that it should provide a categorized directory of useful resources on a subject, Voynich Manuscript research only rarely works like that. For example, some ODP editors wouldn't even link to the Wikipedia page on their subject because it is a deeplink within another resource (i.e. within Wikipedia itself), as opposed to a user's subsite. So only a tiny handful of sites (such as Rene Zandbergen's exemplary pages on the subject, my Voynich News blog, etc) would satisfy a hardcore ODP editor's criteria for inclusion.

In other words, though I hoped that the ODP section might be the right place to bring some kind of order to the disjointed heap of fragments we call Voynich knowledge, many inside the ODP actively argue that that's not what it's for.

But if the ODP is the wrong answer, what is the right answer?

Having just spent a few days tracking down a long succession of dead links, I think there are two main types of webpage that should somehow be actively preserved: research pages (analyses of features, mainly of Voynichese) and theory/hypothesis pages ("my theory that I have, that is to say, which is mine, is mine").

And so I've just started up Voynich Research and Voynich Theories (as subsites on GC's Voynich Central website), each of which I plan to maintain as a single index page giving links to (and brief summaries of) all the research papers and webpages placed there. Though the ODP won't allow links to the Wayback Machine, I can put them in here: though my plan (longer term) is to contact the people involved asking for their permission to put an actual copy on the site, which would be a somewhat better answer.

OK, it's not a perfect answer... but it's a start.