Posts by Doctor Syntax 40432 publicly visible posts • joined 16 Jun 2014
"thus engineers were instructed to build in predetermined breaking points"
It seems we're back to the bad old days of the 1970s. Back then you could buy a fridge and the freezer compartment door would be just about well enough engineered to outlast delivery to the customer. Oddly enough I have a freezer built to those standards today. The only difference now is that you can order spares off the internet.
All you need is sufficient compatibility.
For instance, in a WP document I have settings that ensure a heading is on the same page as the text that follows it. Ditto settings for no widows and orphans.
I'd expect a compatible application to follow that if it opens such a document. And yet I see PDF documents on HMG sites that regularly have a heading as the last line on the page. Either whoever generated the Word file - I expect that's the source - didn't bother with such niceties despite the fact that Word has such settings or their print to PDF application reflowed text to ignore it.
If writers of official documents don't give a stuff about layout why should we worry about 100% compatibility.
If you check through the occasion postings of who contributes to Linux you'll find that "big commercial outfits" are the major contributors.
"Of course, I'm not expecting Linus to be able to magically resolve all this overnight"
He and the other source tree maintainers seem to have been resolving things on an ongoing basis for a long time. The system of having a single maintainer for a source tree has emerged as a very good solution to Brookes' problem of coordinating the work of a multitude of developers. In effect the FOSS maintainer has become the equivalent of architect.
"It's just that the situation makes it more and more inevitable that the Linux kernels we actually run and use day to day will end up in the hands of big business"
The kernels we use every day, as derivatives of the project's kernels, are in the hands of the distro maintainers - some may be big businesses, some not. The whole thing works OK. Red Hat can play all the ricks it wants but the only way they can have a say what ends up in my Devuan kernel is by submitting contributions to Linus, getting them accepted by him, accepted by Greg into the LTS kernel and then into Debian. Even then there may be changes during the life of an LTS kernel. None of this needs copyright reassignments.
"other similar projects - e.g. FreeBSD - don't seem to suffer anything like this level of disagreement"
They're smaller projects, they get less media attention to blow things up out of proportion and the nature of the BSD licence is such that if a big - or small - business wants to take a BSD kernel, modify it and keep the result provide they're welcome to do so because the licence allows them to do so. The GPL doesn't allow that with Linux whatever RHEL may try on.
It sounds as if there needs to be an agreement that it's up to the Rust team to maintain C API changes with some means of coordination to ensure the Rust team have notice and an agreed time period to make their changes so that both can be merged together. Martin might have had a good point but went about it in the wrong way.
Having said that, surely the whole point* of an interface (the I in API!) is that it can remain stable whilst the code that implements it changes. I'd have thought that changing an API means that everything** that uses the API, the Rust wrapper being one will need to change with it.
* All too often forgotten these days, especially with GUIs.
** Depending on the extent of the changes
It says the process will take several months. I read that as the time it takes to do the actual job, not the lead time. Either its a very large data set or possibly it has to be interleaved with other use of the tape drives such as doing the normal backups. I don't envy them.
"To manually check every page before the ad appeared is not viable, and any other option will allow some incidences where this occurs."
Manually? Google are a business making loud noises about how good they are at AI. If they're that good they can check every page on the fly to see if it would be acceptable.
"It will only catch dumb criminals who can't imagine using a product that Apple does not control."
Not quite. The insecurity inherent in any back door makes every innocent users' data accessible to any agency, criminal ot nation state, that finds it. That is what the US discovered only a few weeks ago but what with moving fast and breaking things that lesson has already been forgotten - to the extent that it was ever learned.
"I (and colleagues) work to deliver the will of the secretary of state, no matter how misguided, malformed, or simply stupid their wishes are"
Th HO's core competence seems to be house training Home Secs (apart from the few that don't need it). So effective was it that one of them wrote an account of it in the Times, quite oblivious of what had really happened to him.
Start with Yvette Cooper's.
In fact, she and the rest of the government should publish their own encryption keys, and access credentials for any other online services they use, including banking. And no politicians to use the likes of Whatsapp. Anything else is hypocrisy.
It's not supposed to be implemented if it's not feasible so why would they demand it now.
If they think it is feasible to implement a secure back door all they have to do is commission a proof of concept to be picked over by industry-recognised experts. If they can do that and get approval of the experts then they've proved it is feasible. Until they've demonstrated that it is so it remains infeasible.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
