shellac

What do you mean by 'assignment or creating a new object'? It assigns chat to ... whatever RubyLLM.chat returns. Do you mean the function could have a clearer name?

This has been a thing for quite a while. Thunderbolt is, in part, PCIe over serial (now USB C). There have been GPUs in external boxes, so it is undoubtedly possible, but I don't think they have many users.

Edit: BlackMagic was what I was thinking of https://support.apple.com/en-gb/102103. 'Requires intel processor'

Those who don't remember OpenDoc and Cyberdog might want to get reading.

Most torrent clients I know support UPnP and/or NAT-PMP to request port forwarding from routers. As you say, it just isn't going to work for many people without this.

I run a BitTorrent service for an academic institution, to disseminate research data. We have a regular routable IP address, but still need to navigate the institutional firewall.

I swear this was done deliberately to cause maximum confusion.

Not noted there, but I think use in headlines is restricted to the US (?). This does illustrate the problem with over use.

Worth noting that the author of this letter works at Culham Centre for Fusion Energy, where JET is located. 'Abingdon' was a clue but they could have filled in the detail.

Yes, a minor thing but concur that it is definitely peculiar.

The last table is an awful lot of chrome only bits, many of which are a pretty bad idea on the web.

I'm pretty sure this bot has been operating for much longer than the article suggests (April this year), and truly is a pain. I work in academia and see a lot of ill considered web scraping by ML / AI researchers, but Bytespider is in a league of its own.

Are there free ones? Seemed to be $5.

Thanks very much for clarifying that. It did seem unlikely: I remember NSString (ask you parents...) supported this level of unicode equivalence, and it was quite a burden. Normalising does feel like the only tractable method here, and if you have an extraction pipeline anyway (in rga) maybe it's not so bad.

The (standard) ripgrep regex engine has full unicode support. My reading of that is that it should handle such equivalences like matching the decomposed version.

Ah, it may never have reached the UK. Conversely you've probably never heard of the Oric-1 or Dragon 32. The 8 bit boom was quite a time.

I assumed this would a Spectrum user. How wrong I was: don't remember any Atari machines from that 80s 8-bit generation.

I've done this, but in the other direction, by public transport.

Neither end of the Ridgeway is particularly convenient for public transport. Avebury is quite close the western end, and has buses to Swindon (and then trains to London). It's also well worth visiting: a highlight of the trip. At the eastern end you're not so far from Tring train station.

Accommodation is a bit sparse until you get close to London, so it will tend to dictate how far you walk each day. However there are companies that will arrange it, take your luggage, and pick you up / drop you back on the trail. Some B&Bs can also help with this.

(We just carried our stuff and walked to the places we stayed)

Worked fine on an iPad, or is that not Mobile Safari?

It is mentioned:

> ...Researchers are now working on several next-generation LIGO-type observatories, both on Earth and, in space, the Laser Interferometer Space Antenna;...

It's working out well as a 'forever' language: I can understand and use java code from the dawn of time (aka late 90s, but more realistically 1.5, 20 years ago), and Goetz seems to being doing well keeping that going. The challenge is adding any new features, but they are maintaining a reasonable cadence.

At some point I imagine they will have to remove or at least isolate and disarm some ubiquitous core features (e.g. serialisation and synchronisation).

Ah, I somewhat see the confusion. Green OA doesn't require the publisher to publish other versions for free, it just means you are allowed to publish them. Typically you'd publish them via an institutional repository, preprint server (often discipline specific), or in one of a number of free online services.

> You pay the publisher a large sum of money for the “right” to self-publish the preprint, and they still paywall it

There's no APC with Green OA, so what money are you talking about? Green OA is regular publishing, but with self archiving. There will be a version freely available, and the publishers aren't paid for that privilege.

If you want a route to the death of publishers, green OA is a promising one.

(I think the headline ought to emphasise that this is pushing green OA, which is the interesting bit)

At this point nothing ought to be inserting BOMs in utf-8. It's not recommended, and I think choking on it is reasonable behaviour these days.

I think you have to appreciate just how egregious the behaviour of Microsoft was in the 90s. We're talking about systematic attempts to use their absolutely dominant market advantage to move into new areas and kill rival technologies, plus quite deliberate sabotage of open standards process to prevent the web taking off. (I knew a few people involved in W3C then and Microsoft's behaviour was breathtaking)

Personally I'd avoid comparisons since you're looking at one of the worst situations imaginable, so anything else is bound to seem less harmful.

You find that _less_ scary? Turing-completeness doesn't entail file access etc. etc.

> it would be more appropriate to "apologize" for something that would be considered solely the "fault" of the application

Is this in part a US vs English difference? I used to work with somebody from Texas who was confused and irritated by the english use of 'sorry' for things they weren't responsible for. In english I suppose we use it out of politeness to express some sympathy, rather than necessarily as an apology.

Ditto. Having had to update python code over the years, this looked much easier than python migrations. To be honest moving earlier python 3.x code to current 3.x is comparable, so good job ruby devs.

You're never more than a couple of corners away from the Covered Market in Oxford-land.

Yes, it conflates a structural vs nominal type difference (i.e. punning 'bark'), with a classic variance issue.

Ignore the tree. Suppose you have _seals_, who are plausibly 'barkers' in the same sense as dogs. Take an array Dog[] and pass it to a function that accepts Barker[]. It seems reasonable to take a Barker from the array, nice and safe. But suppose we try to put a Seal into the array, that's not fine: it's a Dog[].

Famously java got this wrong, but runtime checks (somewhat) stop that being a disaster.

> the more posh types in Clifton on top of the hill

In the 70s Clifton wasn't really posh either. Like many cities at the time the affluent had moved out to the suburbs (e.g. across the Downs or Gorge). Difficult to believe now, of course.

(Late reply, sorry)

Well currently you can't get at the AST, only (with some hairy code) bytecode. (The use cases are post-compilation)

But suppose this work did enable that. What Paul is saying is that there are intermediate representations between bytecode and AST that are more helpful for these GPU / SQL / whatever runtime compilers. Representations that capture dataflows, for example. Chance are these compilers would transform an AST to something like this anyway.

See https://mlir.llvm.org/ which is referenced in that talk.