Posts by bazza

Re: Europe

Who knows. There’s been some rumblings of economic retaliation if European countries pass a digital sales tax on companies like Amazon. That’s calmed down so far as I can tell since Joe Biden came to the Presidency. So yes, there’s a chance that the US gov might choose to make an issue of it, depending…

However there’s definitely a mood now in US politics to take on big tech which many perceive to have grown too big for its boots. Google is in the firing line in most states and the federal government, with Apple probably next. [Interestingly MS isn’t, seemingly having successfully given the impression of being friendly to openness.] If this mood persists then Apple’s forced loss of control of browser engines may not be unpopular amongst the US politicians.

What will be interesting is to see how far the politics of consumer rights penetrates into web standards. A lot of what makes web apps rubbish for end users is the ads, bloat, CPU wastage, data slurp, etc with web standards enabling more and more of that. If the web standards were reigned in and enforced, web apps might get better.

But I’m not getting my hopes up, because my experience of straight up ad free web apps on even desktops is that they’re bloated slow junk, for anything other than fairly trivial stuff. Electron - yeurk.

Re: boring old plain 5.18

Who knows. If it does indeed prove controversial, Linus might possibly evict the code from the kernel version he controls. For the time being I'm guessing (and I really don't know, because I've not read widely enough) that it's Intel's version of x64, that's a major target for Linux, and going with wherever Intel's x64 is going is important. Thoughts, anyone?

The unmanned space exploration programme has vastly over delivered. Very impressive!

Knowing what can now be done, one would never send people back to the moon just to do more exploration of it; a rover can almost certainly do just as good a job, for longer and for less money. There has to be another reason to send astronauts somewhere.

Re: Less than 5%!!

Also, whatever the accuracy of what is reported to the SEC doesn't actually matter, so long as it is consistent over a reasonable period of time. What actually matters is profit / loss, revenue / costs. Consistent reporting of fake accounts allows someone to make a comparisons against previous financials, and perhaps then make a reasonable stab at predicting future financials when the fake-account count changes.

It may be that there is no correlation between the fake account count and the financials, in which case it matters not one jot. It may be that there's a good correlation, in which case it's a useful tool as-is. But it could also be a self-fulfilling proficy. If the fake-account estimate is believed, and it goes down, Twitter's advertising revenue might go up due to enthused ad customers. And vice versa. Separating out cause and effect could be a tricky problem, as the only way of knowing is if the ad customers' product sales figures alter in correlation with the amount of ad spend they have on Twitter, but it's unlikely that they can measure that correlation very well.

Changing the methodology, sample size, etc may improve the theoretical accuracy, but there's likely a stage at which further improvement becomes pointless and uninformative. If that stage is reached, and there is still no correlation between financial performance and the improved fake account count estimate, then there's no point reporting it in the first place.

Someone else can probably say whereabouts on that scale of possibilities the current data actually sits.

GPL2 is also vague. It defines "complete source code" as including the scripts to build and install the software, and that these have to be provided in a machine readable form.

Poorly Defined Terms

But no where does GPL2 say that those scripts should be executable, or that they should result in a running program on specific hardware. It also doesn't explain what a "script" is; to most of us it's a bash script, possibly python, a makefile, or similar. But an executable binary that does the same steps, or one of the steps, definitely is not within most people's ideas of what a script is.

It's also perfectly possible to have a complete set of scripts that build the software, but not install it. The installation might be manual (i.e. a human has to copy the image to a chip burning machine, and the chip then has to be soldered down afterwards). They might even have to use a custom machine-specific binary executable to help with this. In this case, it's entirely reasonable that there is no machine executable script for installation, and there can't ever be one that results in a runnable program; a script cannot by itself solder a chip to a board. There can be a readme as a list of instructions for a human to execute to install the software.

Make Up Your Own Device Configuration

GPL2 also leaves room for the following situation. There's bound to be a configuration file specifying device addresses, certain memory locations, etc. This might helpfully get supplied as "set these appropriate to your run-time environment", which makes it hard to figure out what they should be on a specific retail device (no datasheet published). So yes, you can build the software alright, but never know what settings are required for a specific device. And here too GPL2 is inadequate; no where does it say anything about the hardware on which it should be run. The original authors of GPL2 had in mind standards compliant source code compilable on almost any unix-like OS, not specific hardware.

What's also bad news for the vendor is that such configurations can easily change between production runs - i.e. the next uses a slightly different SOC that requires a different configuration - and they may not have records of which SOC was used in a device of any given serial number.

And if one thinks this unlikely, think again; even Apple are constantly churning their choice of components between batches of iPhones, Macs, etc, that outwardly are all supposed to be "identical". At one point they didn't know themselves what had gone into any one specific device, which is why firmware updates to fix issues can have patchy success rates.

In that circumstance the best a vendor could do is supply a whole bunch of configuration files, and tell the recipient distributee that they'll have to munge them together as best they can, see what works...

Now a Contract?

It's this kind of problem that puts companies off using Linux, and why FreeBSD is an attractive alternative. If GPL2 is now also going to get interpretted as a contract, that's simply making Linux less and less attractive, because it is so unclear as to what really does constitute "complete source code", and "installation". Companies might now find they are contractually obliged to facilitate installation on hardware never designed to support post-manufacture installation.

It's also ironic; the Wikipedia article on the GPLs says that they were designed as licenses, not contracts. But now they are being interpreted as contracts.

Here in the UK, a contract is only enforcable with the exchange of money; no money, no contract. That's why you see bust companies being bought for £1 - to seal the deal. So there is a contract if you've bought a TV, but if you've just downloaded an executable for free there is no contract.

This Specific Case

I think the SFC is being casual in its statements in referring to "GPL", and not specific versions. There's a world of difference between 2 and 3. I know the case involves software under both licenses (it's not just the kernel source they're after), but technically speaking there is no such thing as "GPL".

GPL2 does not grant a right of repair; it doesn't say that the re-built executable must installable on specific hardware. It does not say there has to be a USB socket from which an image can be run or burned to on-board flash, or other similar mechanism. However, a side effect of this case might be that the GPL2 is a contract that, in effect, does mandate facilitating third party image installation on the hardware.

Of course, if GPL2 source code is underpinning a product, allowing third-party image installation is actually the polite thing a manufacturer should do anyway. But this case is moving the goal posts in a direction where otherwise blameless manufacturers may now have a deployed fleet of hardware that now is not compliant, and can't be.

So, my worry is that this case will make it even more difficult for manufacturers to be certain that they have a issue-free pathway to GPL2 compliance, even if they're intending on being compliant. If as a result of this case we start seeing device manufacturers move over to, say, FreeBSD, then the ability for people to mod their own devices is actually going to decline, not increase.

And so, what does this mean for the Android mobile phone market? Are we set to see a situation whereby manufacturers are obliged to facilitate third party image compilation, installation, with the expectation of ending up with a working phone? Could be a good thing. Could also be a bad thing, as it means that the third party might not actually be the phone owner. There are good reasons why people might not want a device that anyone can reimage without manufacturer support...

Ambiguity

The SFC said: "Had Vizio produced the source code for the Linux kernel, for the other SmartCast programs at issue, and for the library linking programs, as used on Vizio Smart TVs, a community of software developers would have had the opportunity to modify them to protect user privacy or improve accessibility,"

I hope that in this case, "library linking programs" means software to manage lists of TV programs, and not the source code for ld or whatever the object code linker that is used.

Re: Insight?

Of course the issue of dust accumulation was thought about by the development team, and allowed for. That's why it has lasted nearly 1.5 earth years longer than the planned 2 earth year mission duration.

Re: More reason to relocate server farms

Brings a whole new meaning to the phrase Garbage In, Garbage Out...

That "no code change" is something that I treasure today, and go to some lengths to replicate if needs be.

I quite like ZeroMQ for its diversity of sockets, ranging from in-process to cross network. Just change the connection strings, no other changes. If it implemented CSP instead of just the actor model, it'd be perfect.

One of the pities about things like OpenMPI and OpenMP is that they don't play nicely. If you parallelise an algorithm using OpenMP you then can't easily scale up with OpenMPI.

Re: Even if they did...

Battleground or not, it would cost MS money to resource an OSS and maintained version of NTFS, and there is simply not the business there to warrant it.

A real question is how much maintenance does the formerly Paragon code actually need?

"Raytheon World Peace Missiles"

Well, if you have enough of them and then never have to fire them, they kinda are World Peace Missiles.

(sales brochure extract)...

Er, Javascript does threads?

This kind of fault is caused by not dealing with error conditions thrown or returned by code. Someone somewhere assumed "That error will never occur". So, where else have they screwed up in a similar manner?

Re: Hang on a mo ...

Except that, where MS's driver interfaces are generally very stable - an aid to the independent developer - things are somewhat unpredictable in Linux. That does not help the independent developer.

Sounds like it, though I think it's interesting to see where their effort is spent. Server type things eg SQL Server, but not desktop like things such as Office. The year of the Linux Desktop has not arrived.

Re: SpaceX is his most solid buisness at this point.

SpaceX is not solid. Falcon 9 works and has customers, but Starship is not finished. Only when it is finished, and only when they have Starlink earning full revenue will the company itself be solid.

Broken a rule, or broken a law?

Either way, the lack disclosure about his acquisition is undeniable, and is going to get him into some sort of trouble. The only real question is, how much? Furthermore, there's now at least two federal authorities unhappy with his attitude (the FAA, SEC). At some point the authorities are going to have to teach him a lesson. I'd be intersted to know just how big a book the SEC can theoreticaly throw at him for this one. A fine? Jail? Disqualification?

Another thing worth mulling is, where would the $43billion come from. He's not got that as cash. For Twitter shareholders, I can't see the attraction of a stock swap for Tesla shares. He'd still be in an influential position over their value, and they wouldn't be. And if they all decided to bail out of Tesla shares ASAP the current price would drop. That sounds like an unappealing lose-lose situation, so I don't see why they'd accept that sort of offer.

Re: "Auctioneer puts Space Shuttle CPUs under the hammer"

There's always something missing from the requirements set...

Re: LibreOffice, for example, is every bit as good as Microsoft Office

I've previously found LibreOffice's spreadsheet to be slow compared to Excel when handling large amounts of data. Excel is just better.

LibreOffice is pretty good for the resources that have gone into it. But MS have had a large number of developers working on Office for decades, and it shows.

Re: It's not clear what the equipment is,

(though the cat may or may not be there either, but regardless it will be angry)

Re: Holy Microsoft?

I didn't see it, but I recognise the smell of it. I've not tried 11 yet, but I strongly suspect that you'll be proved correct, judging by everone's "Meh" (at best) response!

Re: Workforce Demographics

Except, according to this article, it isn't quite the case that there is a net equivalency in how Mac and Windows can be administered. This article says that:

"The upshot is that on a JumpCloud-managed system, the user will be gently nudged (with a soon to be configurable message) that, maybe it's time to get those updates done?"

I don't know whether that limitation of JumpCloud is universal across all third party Mac enterprise admin tools, but I see no reason to suppose that any other third party product isn't similarly limited (all corrections welcome!).

I think that that specific aspect matters a lot, especially if you're following any kind of zero-trust security principles. Taking that philosphy to the limit, you'd not be letting a badly out of date un-unpatched machine on to the company network or services. It's the user's fault for not updating as and when prompted, but when it's finally considered too toxic to handle it become the admin's job to sort it out.

I absolutely agree that one's mileage will vary. There are of course some users that can be trusted with keeping things as they properly should be. Then again, there's an awful lot more who'd rather not have that burden added to their lives, and a few who absolutely should not be given any say over corporate IT admin whatsoever. If Apple really are saying that the user has to be in charge, that just doesn't suit a large number of people.

Yes, imaging is still a thing. There's not a lot that Windows won't deal with if it finds drivers missing. That means you're not having to maintain the stock company image quite so much to account for the myriad hardware, so it can be nothing more than a quick way of dropping in a guaranteed dross-free Windows install.

I wouldn't be surprised if Apple do take a dive into Enterprise management, but I'd bet that they'd do it as "Apple administer this machine as a service for your company", instead of giving companies the freedom to admin as they see fit. From Apple's point of view this would be fantastic - they'd displace all the third party products that are currently in-use, they get a revenue stream, and they can extend their walled garden even further. It might even suit companies, who can then dispense with their admin staff. And if problems arise, well that's unlikely to annoy enthusiastic embracers of Apple's shiny offerings.

Re: Risc V the Chip Equivalent of Linux?

It's not millions, it can be as cheap as £50k to have a chip made on an old fab. You need to have the masks already for that price, but it can (and is) done by specialist's working in the component obsolescence business.

Depending on what you're looking for it can be cheaper to get a chip made than to muck around with FPGAs...

That's still not going to be cheap enough for a determined fab to be outweighed by those very few who would go their own way.

Re: Security is for the weak!

Dirt

Perhaps as part of ensuring that it remains that way, protect it from idiot investors out to make a ton of profit instead of running the business.

There is a veritable fuck ton of companies, jobs, software that depend on Arm being preserved more or less as it is today. One thing worse than it being taken over by your deadliest competitor is it disappearing altogether through financial overgearing and unrealistic owner expectations. A consortium could insulate it from that and if properly constituted could address competition concerns too.

Re: Speed up shutdowns?

There's nothing in the AC's post to suggest that those things are hosted on singular hardware.

The order makes sense; there has to be some DNS up before AD will work. The AD has to be up at least somewhere before the LDAP can kick in (I'm making assumptions about their LDAP and what's hosting it). And in a zero trust setting, you'd want security and logging up from the get go.

Zero Trust makes start up orders interesting.

Re: really cemented it though was the US DoD, in the 1980s

DoD didn't mandate Ada until 1991, but granted a ton of exceptions. Not many projects actually went ahead with Ada, being based on C / Posix instead. The "Ada" mandate lasted 6 years only.

In 1988 POSIX effectively defined the standard against which one had to write C code to be portable as per DoD's requirements (which is what the DoD actually wanted). By 1992 POSIX was also specifying command shell and command line utilities. POSIX and The Single Unix Specification merged in 2001. At the end of the 1980s DoD also mandated open hardware standards for various classes of system, plumping for VME (which is still supported to this day). ADA, where it was used, had to fit in on top of all this.

It depends on what you mean by "Unix". These days, something can officially be labelled "UNIX V7 (tm)" if it passes the Open Group's compatibility tests against POSIX:2008 2013 edition, aka Single Unix Specification v4. Note that OSes can meet these requirements even if their native shells, APIs, etc. are not unix-like in anyway whatsoever.

Pretty sure Linus did Linux for fun / university, originally, not because he was passionate about the corporate doings of AT&T. FreeBSD was the open re-write project that was motivated to wrest "Unix" out of proprietary control. It came out at about the same time as Linux. Both ended up achieving the same end result.

Re: Nothing new, kinda pathetic really

Yep, that is what will eventually happen. Probably worth brushing up the old C...