Posts by haiku 25 publicly visible posts • joined 17 Jul 2014
To quote from an article headed "Android’s toothless “Privacy Sandbox” fails to answer iOS tracking limits" in this morning's Ars Technica:
"Android "Privacy Sandbox" is optional for advertisers, and that misses the point"
As is well known, if it's optional it don't count.
Note to Reg's editors: hit the coffee a little earlier guys & gals :)
II have always believed that many of the problems experienced with today's software are much the same as they were yesterday, and the day before that i.e. simple bad coding practices.
This is born out by the stats maintained by OWASP: when "Injection" is (after several years) still one of the top three security flaws, we have a problem that won't be cured by, for example, block-chain - said block-chain being IMHO much the cure for which there is no known disease (with thanks - and apologies - to the late, great Victor Borge).
See https://owasp.org/Top10/
http://www.computinghistory.org.uk/det/1788/Don-Estridge/
Don was the father of the IBM PC, and it is due solely to Don that the PC was launched with a totally open hardware architecture that allowed companies such as AST to produce cards such as the "6-Pack" (so-called since, on a single card, it included serial ports, parallel ports, additional memory etc. - available as discrete components from IBM).
Don also decreed the totally open software architecture: a program listing of the code was included in the Technical Manual. Which led to the situation - since the code (and the bugs in the code) were copyright - wherein the clone manufacturers had to emulate the bugs (for compatibility) without actually copying the exact code ... :)
Don was way ahead of his time, especially where IBM & the PC were concerned, He realised that the IBM clients preferred to purchase from a single supplier - especially if one had a "valued client" (or some such) licencing agreement - and so were not interested in products from Apple but were rather waiting for a similar product from IBM.
So, so the story goes, he nagged the IBM board until they reputedly allowed him to manufacture the PC, if nothing other than him quiet.
The rest, as they say, is history.
He unfortunately died in an 'plane crash, just as his vision was bearing fruit. One wonders what the PC world would be like today if he were still alive.
"
The other is the self-perpetuating monster ISO9000 et al. Originally supposed to bring the benefits of Far Eastern continuous improvement quality systems to Europe & the US it has had the exact opposite effect. Provided you have documentation that you've followed your 'process' you get - for a fat fee to a B-ark consultant - a certificate that you're compliant. It doesn't matter how good or bad your process is or how inappropriate for your business provided it's documented & followed. The bureaucratic cost of changing your process means improvements are positively discouraged.
"
I was a developer at a company that decided that ISO 9000 certification was worthwhile.
My take, at the end of the process (during which we were successfully certified) was that as long a the SNAFU's were documented, you would be certified.
Not that there was any need to do anything about said SNAFU's ...
And, as for the certifying consultants - academics all - the less said the better !
In the pre-PC daze of yore I once tried to report a compiler error for an HP3000:
HP: Sir, you cannot report a compiler error via the telephone.
Me: Why not ?
HP: Your bronze support level does not allow telephone call support !
Me: I am trying to do you guys a favour ...
HP: I am sorry sir ...
Me: Ah well, I will have to speak to the company's lawyers about being sold defective goods.
HP: Defective goods ??
Me: Yup, the compiler that sold us has bugs and is not fit for purpose.
HP: Every compiler has a bug or two.
Me: I know that and you know that, but will the judge know that ?
They accepted my bug report via the telephone ... :)
My first "eureka" moment - understanding the concept of McLuhan's global village - came when posting a question (sitting in South Africa) on CompuServe (in Ohio) and receiving an answer (from Holland) ten minutes later.
For some or another reason this blew my mind, even though I had used FIDOnet for many a year.
How much work is this going to be, really? Is there only one question - "The Question"?. Or is it one per week, day, hour, phase o' the moon? The "Time Limit" paragraph hints at "Questions"?!"
Exactly my first thought.
Then I realised that the actual challenge was in fact more life-like: make sense of an ambiguous specification ... !
IT was a cost centre in the company that I worked for in the late 70's, with all IT expenses reflecting on the companies' monthly financial statements.
All work for the other companies in the group was charged for on a fixed-price contract basis, with the other companies having the option of employing outsiders. So pencils were pretty sharp when it came to both quoting and delivering.
The first benefit of the system was that all companies - as part of the annual budgeting process - had to budget for their IT requirements, so we had a pretty good idea of what was coming, and could plan accordingly.
More importantly, however, was the fact that all work was going to reflect on somebody's budget. So the 'customers' made sure that everything was specified correctly 'cos, if not, they were going to blow their budget on re-work. Ditto the IT department if we screwed on a development.
And asking the Group Financial Director for extra funds was, to put it mildly, not something one did twice ... 8)
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
