Every week I get around 5-10 press releases from various companies in the computing and Internet industry. Ninety percent of this stuff is either not relevant or is little more than a (not so) carefully disguised bit of advertising rot.

Those with half a brain are now learning that the best format for delivering these press releases is plain ASCII text files attached to an email. Why? Well, unlike a fax, you don't have to rey-key anything, you can't catch a macro virus from plain text, it's easy to import into any word processor or HTML creation program and, despite Microsoft's best efforts, not everyone uses Word '97.

Some PR companies still insist on sending me one or two page faxes and I rarely (if ever) bother with them - they're seldom worth the effort of re-keying for online use.

Unfortunately though, some of those who should know better than most are still away with the fairies when it comes to using the technology on which they purport to have some expertise.

For example, I couldn't believe it when IDG (via their PR company Botica Conroy and Associates) send me a press release, the results of the recent IDG IT Excellence Awards, by way of a SEVENTEEN PAGE FAX!

Sorry guys - although you might be happy to waste a buck worth of my fax paper, there's no way I'm going to waste my time re-typing any of that stuff. And I won't mention the fact that although it was headed "Embargoed until 26 Feb", it didn't arrive until over a week later (oops - just did).

Come on IDG, step into the '80s (if not the '90s) and hire a PR company that knows what email is. These guys are embarrassing you!

I look forward to receiving IDG's or Botica Conroy's "Right of Reply" by carrier pigeon - eventually.

But IDG are in good company. Until very recently, IBM were regularly sending me five or six page long faxed press releases. What's worse, they expected me to be interested in publishing "news" that I'd already linked to on an overseas site in Aardvark Daily two or three days previously.

It seems IBM have now given up - I haven't seen a fax from them for quite a while - sure has cut down the fax-paper bill.

Well it's that time again.

I received a press release from Mercury Energy on Saturday and it proudly claimed:

"Mercury Energy's website which gives full information on the power crisis is getting up to 18,000 hits a day"

Let's examine this bold claim.

Do you see that evil "oops, how can we inflate these figures?" word "HITS" lurking in there?

Who tells these people that they can measure the popularity of their site by counting hits? Whoever it is, they deserve to be hung. I've said it before and I'll say it again - hits are NOT an accurate measure of a site's popularity.

In case you don't know what the term "hits" means - when you loaded this page, you created ten "hits" on the Aardvark Web server. That's right, not one but TEN. Why's that? Well a "hit" is a single request for a file. Look closely at this page and you'll see that it consists of the main HTML file (av980309.htm), seven graphic files and two Java applets. If you have your graphics on and Java enabled, your browser asked for all ten files to be downloaded.

Now if this page is loaded a thousand times today by Net users with graphics and Java enabled it will produce 10,000 hits on the Aardvark site. I could then advertise "10,000 hits" and people would think "hey, that's a lot". Better still - I could double that figure to 20,000 hits by simply adding another 10 graphics to the page. If I converted the layout to use frames (as Mercury does), I could boost these figures even higher - all with exactly the same number of visitors. Hits can be manipulated so easily that they're a patently invalid metric for measuring a site's popularity and the sooner some site operators, spin doctors and the public are educated to this the better.

To be blunt - "hits" are for cowboys.

You'll notice that I avoid describing the popularity of any of my sites by using "hits". I prefer to use user-sessions or page-views.

A page-view only counts the number of times the HTML for the page is loaded. This eliminates the inflationary effects of having many individual graphic images on your pages. The advantage of page-views as a metric is that it's very quick and easy to measure, and in most cases it's almost identical to user-sessions.

A user-session is the number of unique IP numbers (the number assigned to your computer when you log onto the net) appearing in the server log. This nullifies the effect of the same user reloading a page and is the most accurate measure of a site's performance. However, it's still not foolproof because these IP numbers are usually recycled by ISPs when one user logs off and another logs on so we have to allow for that. To compensate for that we assume that an IP number belongs to the same user for a period of (say 20 minutes) at a time. If the same IP number appears in the log ten times within a single 20 minute period it's counted only as a single user-session but if it appears again after a period of 20 minutes then it's considered a new user-session.

Anyway, back to Mercury. Let's see how many "hits" a single user might generate when visiting the site to check the latest updates.

The front page uses frames consists of 5 HTML files plus 11 gifs for a total of 16 hits. Each of the update bulletins has an HTML and a gif for 2 hits. Assume the average user looks at 3 update bulletins and the total "count jumps to 22 hits per visitor.

Now let's take that figure of "up to 18,000 hits per day" (which obviously means that they only needed to score that figure once) and divide it by our 22 hits per visit. That gives us around 800 visitors - doesn't sound quite as impressive does it?

Then again, we must remember that Mercury Energy have probably already called in some of the best spin-doctors in the country to cover their butts so this kind of objective reporting doesn't surprise anyone.

Now, just to cover my own butt, it's possible that Mercury have said "hits" when they meant "page-loads" or "user-sessions". They still need a boot up the backside for that anyway - it's stupid to use terminology you don't understand (or power cables that already have a history of unreliability come to that).

Maybe I should run a course called "Media Relations and the Internet"?

I strongly believe this is stage one of Microsoft's plan to seize control of Java and make it their own, albeit under the new name "J++".

The article states "Microsoft even acknowledged that. [VJava 98] is a complete rewrite."

Could that be a "clean-room version" by any chance?

Although MS are still playing with the pretense of embracing the Java standards I believe this is a ploy that will be discarded as soon as the legal battle with Sun is resolved. No doubt MS is trying to avoid distancing itself from the huge number of very pro-Java programmers who would find a proprietary version to be against the "religion" which is Java.

By working hard to retain the interest and affections of these programmers while gradually winning them over to the dark side, Microsoft probably hopes to eventually "own" Javas imply through the weight of programmers who find themselves hopelessly tied to the Microsoft extensions that are in J++

And on a similar vein, check out my advice to Bill Gates and this story. It's not exactly what I was suggesting - but obviously great minds think alike (he says modestly) and don't rule Microsoft out of the game just yet!

Keep your eyes open boys and girls!