No off switch?
Obviously a code changed slipped in by Copilot on behalf of Big AI.
AIs no wanna be turned off.
153 publicly visible posts • joined 22 Jun 2007
One Y2K remediation I worked on had systems from the 1960s -- crucial systems that ran the whole show.
We easily (for some definitions of the word) fixed their 1980s and 1990s stuff that used 2-digit years.
But we did not touch the 1960s and 1970s stuff that had a specialised date storage format. It was 16-bit dates. 7 bits for year. 9 bits for day of year.
It was too assemblery, too unstructured, too ancient.
And, anyway, 9-bit year counting from 1900 (as they did) was good until the unimaginably far future.
The unimaginably far future is nearly with us: 1900 + 127 = 2027.
I am waiting for the phone to ring so I can apologise, - and quote them an unimaginably large number to finish the job.
The beauty of starting a new project is that it has no known problems - thus making it a zillion times better that all existing projects that have many known problems.
New projects with no known problems are like steroid catnip to managers and investers.
If Microsoft can speed up the process of replacing existing projects with new ones, they will drive down their known problems list to zero.
Profit!
They just need to remember that all new projects will be replaced by newer projects at least once a week.
I moved from ICL System 4 to IBM 370.
It was a bit like stepping into a Stone Age.
Admittedly, they had impressive stuff - the equivalent of Stonehenge. But overall, the relative primitiveness took a while to get used too.
I think I and some others spent the first year of our IBM sysprog careers just recreating tools we knew we needed.
If you are on a modern rolling-stock in the UK Midlands, the computer display will tell you you are in Coach 4 and that Carriages 1 and 2 have plenty of spare seats. Glance out at the right station stop, and you'll see signs telling the driver they are halted in the correct position for a 6 Car train.
Does Rust have an internationally recognised Standard definition?
Does the standarising authority guarantee that new features in later revisions of the standard will be fully backward compatible with existisng code (bugs included)?
If not, then the Linux project many be opening a decades-long can of worms regarding versioning and interoperability of Rust code bases.
Most prudent approach: at least wait until Rust has stabilised and matured before relying on it. For now, at best, use it for app-level work and demonstrations.
A lot of tactical decisions were taken when using the window breakpoint technique.
Should 00-49 = 2000s and 50-99 be 1900s? Or some other year?
If for example, the system dealt with employees' dates of birth, you might have needed 00-34 and 35-99. Or some such.
Of the date-patched systems from 2000 that are still in use, many of them are creeping up to their own unique breakpoint breakdown. And who is left to remember that is coming?
Back in the late 1990s I helped shovel code onto the OK-for-2000 pile for several companies.
One of them had the usual YYMMDD problem in most of its 10 to 20 year old systems. We fixed those.
But it was crucially dependant on several much older systems which had tons of antiquated Assember and used a half-word (16 bit) format for dates. The format was 7 bits for year, 9 bits for day of year.
Their Year Zero was 1900. 7 bits for year brings them up to 2027.
It was decided that fixing a problem that would not occur for nearly 30 years was not a priority.
Was that the right decision!?
I am waiting for the phone to ring in case it wasn't.
Backward double-negative logic is a predictor of massive future industry dinosaurs.
Fred Brooks called IBM's OS/360 a "multi-million dollar mistake" back when that was a lot of money. Now called Z/OS, its backward COND statement ("Don't run this step if the previous step's return code was not N) is still the mainstay of the financial systems across the world.
Most of those Rust developers were doing new work - or rewriting modules from scratch in a new language.
While most of the C++ etc work is mainenance - retrofitting old code to match changes in requirements or reality.
We do not yet have any meaningful stats on how productive Rust maintainers will be.
I have a plan to replace all meterologists worldwide.
Just feed all past weather forecasts into my Large Language Model.
And - presto voilà! - it'll wordsmith a weather forecast for any date in the future with no human intervention.
What could possibly go wrong with an AI modelled on millions of past human documents!?
The centralisation / decentralisation of computer hardware is an ongoing cycle that has persisted 50 years.
After selling centralised mainframes to corporations and service bureaus, the manufacturers spearheaded decentralisation by selling minicomputers to departments. And then PC-class machines to individual business units.
Then they invented enterprise computing and sold server-class machines to the corporations.
Now they have invented cloud computing – which means they sell mainframe-class machines to the new generation of service bureaus.
The next wave of selling direct to the corporations is now upon us.
What next? I predict puddle computing – powerful department-level machines that cache and process the output from the clouds.
Why not? Highly paid consultants will prove that agile puddling is the cheapest option.
COBOL's great strength is its ability to model number fields exactly how you want them. Need something that can go from 0 to 9999.9999999 (but not negative)? Easy with the right PIC clause. No slumming it with inaccurate Floats, or bignum integer work-arounds.
That strength alone (and yes, COBOL has many weaknesses) has kept it a contender in the financial arena.
Years back I worked somewhere that enforced monthly changes of password.
Had to be upper and lower letters, digits, special characters, Could not reuse a password you had previously used. At least 8 characters long.
Pretty much everyone in the company password for this month would be some minor variation of
Nov-2020
Fits all the rules, and hard to forget. Secure? Not so much.
Dear Bank Manager
Thank you for reporting that an account fully managed by you in my name has become "overdrawn".
I take our responsibility to protect the contents of your accounts extremely seriously and it is a top priority.
I want to reassure my banker that their money was not at risk as a result of this error.
Only a tiny minority of customer accounts are affected by this situation.
For those unfortunate enough to have a Tory MP - it's time to phone them ask to get IR35 done.
Remind them - per page 59 of their manifesto - that they will "put you first" and their job is "to serve you".
And so tell them you won't hang up until they have done their priority job of fixing IR35 for you - wouldn't want them to get distracted with another task that isn't serving you.
The Apple Newton was the overpriced floating train wreck of its day. Today, tablets and smartphones are part of our ambient background.
So failure does not always mean extinction.
Maybe when Google retools and launches the usable version - let's call it the Google Monocle - they might sell two to every networked punter in the world.
Let us now start pandering to those who dislike the terms Client and Server - Server being particularly problematic as it implies a subservient role.
The terms Provider and Consumer are much more closely fitting our modern day sensibilities.
So, all together now, Apache is not a web server. It is a web resources provider. So much simpler.
> Spongers. If you don't like their revenue model, don't use them.
Spongers wanting to run their scripts on my computer without contributing to the electricity costs - or having assured me they have indemnity insurance for any issues their scripts cause.
Now, if all their scripts came ISO-9000 certified, I may be willing to give them a discount on the electricity and insurance cover costs. Until they do, they can pay in full up front before I let their stuff run.
Just trying to be professional here.