Search Engine Optimisation: Art or Science?

I'm not an expert on Search Engine Optimisation (SEO) and leave this to people who know far more about this subject than myself. But in one conversation on this subject with a few Internet industry people recently, we ended up discussing whether SEO is an art or a science, and I thought this was a great topic to mention further.
Now I've previously said (in various forums and presentations) that online marketing itself is more a science than an art, but this has been based more around:

1. writing an online marketing plan/strategy
2. implementing it
3. analysing it
4. learning from it
5. refining it
6. etc.

However the work of an SEO person is less easy to test as:

• You don't get immediate results (spiders take time to crawl)
• Variable amounts of content are produced about each term (making organic search results varied)
  New content is constantly produced
• The exact ranking/prioritisation of search results are closely guarded secrets
• The algorithms behind these results can change over time
• Each search engine has different algorithms(and a bunch of other more technical things that get increasingly complex the more you ask on the subject)

No wonder you need a PhD to understand this stuff!

From this lack of immedate results springs a set of businesses that are focused on spreading theory, rumour and suggestions, then testing them to destruction... the SEO industry. This is hardly a scientific community and one where discussion boards are regularly set alight by the merest hint of actual information.

For clients who use these services though, SEO agencies can perform wonderful feats of illusion and in some cases magic (one apparently even made BMW's German site disappear from Google for a while back in 2006).

But hang on, there may be a glimmering light out there in Optimisation Land. Some people are actually trying to put the science back into search. For example Professor Mike Thewell (who has the wonderful title of "Professor of Information Science and leader of the Statistical Cybermetrics Research Group at the University of Wolverhampton" - didn't Doctor Who face him back in the 1970's?) has recently started blogging over at http://linkrelationships.com/ on exactly this topic.

His intial post on Scientific SEO sums things up nicely:

But at the end of the day decisions are based upon gut feel and received wisdom:
art rather than science

His research & insight (which can be brain-hurting to mere mortals) should be interesting to watch over time, as he divulges more about his research. If academics are to eventually try and add credibility to what was previously just subjectivity and conjecture, surely that cannot be a bad thing?

My verdict... For now, lets call SEO work a 'craft'. Something that is on its way to absolute definition and measurement, but still shrouded in arcane terms such as 'black hat' and spiders.

(I think it must be Halloween soon)