Tuesday, May 04, 2004

It strikes me that anyone buying into the Google float needs their head checking...

Having said that, it'll probably rocket 200% and make a fortune for everyone. Apparently non-Americans aren't invited to the party, so I don't have to choose.

One interesting comparison the Google owners make, in justifying their split voting structure, is with Dow Jones. I worked in financial IT in the 90s, and followed DJ's progress quite closely. They acquired a trader information subsidiary, Telerate, that competed with Reuters and (later on) Bloomberg. Reuters kept their product up to date and consequently made large amounts of money (until the web came along and they tripped up badly). Telerate stuck with what was basically early 80s technology until pretty much the end, relying on their newswire to keep up market share among the minority of users that *had* to have their service. Eventually, the family (holders of the voting stock) got fed up and made them sell Telerate (not for very much money, as I recall). I'm not sure that this is a glowing example of the merits of minority control.

Didn't Hollinger have split stock as well?

The other interesting statement is that the A and B shares have equal "economic" value. Presumably (I haven't read the whole prospectus) the A shares will one day belong to Brin & co's heirs and successors. I'd think that if a takeover situation emerges in many years time, the grandkids (or charitable trust, whatever) that owns the A shares are going to be able to demand a substantial premium for their voting rights. I'd call that economic value.


The auction process, as advertised in the prospectus, claims to be designed to reduce price volatility post-float.

As a rational investor (assuming I actually wanted any of the stock...) I'd prefer to buy in the market (where I'd know the price and how much stock I'm getting) than participate in the float.

So the only reason to buy in this, or any, floation is to get the stock cheap, either to stag the issue or to acquire a holding cheaper than I would in the aftermarket. If Google have designed a system that will prevent the stock opening at a profit, then why take the risk of joining the auction for nil reward?

Of course, if the market realises this, the auction will be less popular, so the stock will be cheaper...

I'd agree with Ideoblog on sticking to index funds

Tuesday, April 13, 2004

The BBC has a report (with comments) on Google's proposed GMail service.

Interestingly, the (mostly UK sourced) comments (from "ordinary users") are mainly in favour. There is one person who objects to how Google already invades his privacy by reporting everything he posts on public newsgroups and discussion lists. Hmmmm.

As most people said, if you don't regard targetted ads as a reasonable tradeoff for a free 1GB, then don't use the service! If your mail is highly sensitive, then run your own mailserver, use PGP or SMIME, code your messages on a one-time pad and send them by carrier pigeon, whatever...

The cynical view, of course, is that Google don't actually plan to go live with this, at least, not until disk prices fall by another factor of ten. The idea is simply to stir up some publicity ahead of the IPO, show that they're still "innovative" and then can the idea "in response to public concerns".

I'd certainly use the service - I need somewhere to redirect list subscriptions, registrations, etc and Hotmail's limits make this increasingly irritating.

Monday, March 22, 2004

Casino Capers documented at http://www.guardian.co.uk/crime/article/0,2763,1175810,00.html.

Someone appears to have taken the Ritz Casino for a bit of a ride using physics to predict (roughly) where the ball will fall. The plods are considering prosecutions, apparently, but there's some doubt as to whether any laws have been broken. I don't thing any should be - why should the casino get special legal protection if their randomness based games turn out to be less than random?

Thursday, February 26, 2004

Error Message: Your Password Must Be at Least 18770 Characters and Cannot Repeat Any of Your Previous 30689 Passwords

Apparently this is caused by a bug in Win2k (KB 276304)

I'm sure there are numerous sysadmins who would consider this a reasonable policy, however..

Monday, February 02, 2004

John Naughton wrote an article in the (UK) Observer last weekend, criticising, as one might expect, Microsoft, Bill Gates and Gordon Brown's alleged adulation of the latter.

An interesting aspect (to me) was his characterisation of Gates as akin to a "rebellious teenager". Certainly Microsoft is (still?) run very differently to other corporations. I did some (very brief) research on the leadership of two other substantial technology companies, IBM and HP.

HP is led, no doubt effectively, by Carly Fiorina, a medieval history graduate and Honorary Fellow of the London Business School. IBM's current CEO is Samuel J. Palmisano, a 30 year veteran of Big Blue management. Very clever people no doubt - probably very personable and incisive decision makers. However, I would lay large bets that if you asked them to explain (unbriefed) the benefits of a 64-bit CPU architecture, or Java over C++, or the likely future direction of human-computer interfaces, they'd struggle. However, if you asked Bill Gates, he could, I suspect, give you all the detail you wanted and then some.

This leads to a different style of business. At Microsoft, Gates and Ballmer still make most of the big (and even medium) decisions. In IBM, HP and 95% of other technology companies, everything gets devolved up and down the management tree until some sort of consensus emerges. They hire pretty competent staff and managers, who produce solid-but-dull products (like Compaq servers) and steady if unspectacular growth. The "growth engine" of these businesses is really corporate finance - making the right calls on who to merge with.

Companies like IBM and HP are (these days) not likely to offend public opinion. Those companies (such as Microsoft) that are run by "rebellious teenagers" probably will - does this mean that politicians like Gordon Brown should stay away from them? Not in my view - these firms are going to be the ones generating growth and innovation and politics needs to engage with that.

Wednesday, January 28, 2004

Much comment around on Sir Williams promise to "end Spam in two years"
Oddly, I can't find the actual speech on the MSFT website - perhaps it's been erased by those pesky spammers?

Do I think it's credible? Possibly. Right now, state of the art anti-spam tools run a hit rate around 94%. Brightmail does better than this (I think) with a real-time blacklist of mails sent to one of their "honeypot" addresses. We're seeing a bit of a technology race as the spamsters work around filters and then get found out and re-blocked by each incremental move in the technology.

The limiting factor is mostly false positives - I could probably reconfigure filters to drop 99% of spam on the floor - but I'd lose a few messages (like circular emails from long-lost mates, or booking confirmations from airlines) that I'd quite like to see. So what Gates (allegedly - as I say I can't find the article) talked about - cash at risk schemes - are going to become important. (as I understand this, this means that the mail is digitally marked to say that the sender will pay you $XXX if the mail turns out to be unsolicited bulk mail - this gives the message a free pass through spam filters and ensures - in theory - that you don't miss out on it).

So spammers hit rates are going to fall. Currently, if around 30% of people run a "good" anti-spam tool (e.g. 90% efficient), 73% of spams will be getting through. If in 2 years, 95% of people have a 95% effective system, that'll drop to just over 5%. Also, anti-spam laws are going to do what they do best - constrain the law-abiding. So "real" businesses (like Dell or whoever, maybe even Indian outsourcing companies) will start to be a lot more careful about how they do online promotion.

I guess that leaves the fraudsters and pornstars. They probably won't be restrained by antispam laws - after all hacking into someone elses computer is, rather sadly, fairly illegal in most countries. Getting rid of those guys completely is going to be down to better security, compartmentalised systems, etc - which won't be widespread in two years.

My best forecast would be that 2 years from now, spam is going to be a lot rarer - one or two mails a week, not four or five a day. 10 years down the track, it'll possibly have gone the way of the Moa (as will viruses). Of course, new techno-irritations will take their place (does anyone remember the 640K MSDOS memory limit or the Windows 3.x handle limits?)

Monday, January 19, 2004

Computerworld has an article by Scott Berinato The Future of Security.

Berinato talks of "a digital Pearl Harbour". (Digression: I find it interesting that Pearl Harbour still resonates for the people of the USA after over sixty years. Clearly, it had been traditional in warfare up to that point for an aggressor to give full and specific warning of an intended strike. For the Allies though, it was rather fortunate that the precedent set by Tojo absolved Ike from giving the Wehrmacht full details of his D-Day plans)

The scenario painted in the article involves a computer "attack" leading to societal breakdown, death, destruction and primetime TV cancellations. It gives an interesting insight into the securo-geek mindset (possibly the far extreme of that mindset - but not dissimilar to the people that brought your the New Zealand Y2K fridge magnet).

The article suggests a "global and instantaneous" attack on the Internet's infrastructure, taking down routers, domain servers and "industrial systems connected to the Internet, like utility control systems". Taking down the Internet is, I guess, possible, even likely. However, there are precious few utility control systems even connected to the 'net, let alone which use it as a prime means of communication (if they did, the lights would be going out on a regular basis every time Star Trek downloads crowded out all the bandwidth).

Berinato then posits a failure of the bank networks, leading to mass starvation as cash runs out and people are unable to buy food. Firstly, this assumes that the virus managed to cross banking firewalls (believe me, the backend machines that keep your account balance are kept very isolated from the Internet). Secondly, there are alternatives to computers: my (UK) chequebook has a calendar printed on the inside back cover. I can go in any branch, show some ID and get a couple of hundred pounds out against a cheque. The cashier makes a hole in the date with a biro to stop me getting more cash the same day. Simple - and no computer needed. Even with all their computers out, the banks would just need to hoard the cheques until the systems got sorted.

Anyway, sticking with the scenario, the virus somehow spreads into the electricity network and shuts the power down. "Then the lights wink out. Everywhere. And it begins to get cold". People start freezing to death. Actually, very few people live in places where life depends on continued electrical power. (Canada, maybe). I've slept in a tent in -10C - perfectly survivable, just put all the clothes you own on and you'll be fine!

So, the article concludes that post the "attack", the government (there is only one, of course) will panic and restrict the way we use computers. No more Excel macros, Flash websites, or even forwarding email messages (you type it in again, boy!). I guess that's plausible, the US government does many stupid things in knee jerk reaction to events - fortunately most of us don't live in the US!

I'm not sure if the author really believes in his picture. He seems keen on "zero tolerance for insecure software". I think this would be a pity - I quite like my insecure software..

This page is powered by Blogger. Isn't yours?