Posts by JLV

Re: Homebrew is a disaster

I am surprised at so much anti-Brew sentiment here. Anyone care to explain?

Honest question. Not defending Brew, my goto is MacPorts.

The reason I am asking is that Brew-love is pervasive on macs. Go to github and, 90% of the time, you’ll only see `brew install` as instructions. Nothing about `ports`. Devs typically only talk if brew as well. I’ve learned to just auto translate brew -> ports mentally myself.

Hence my surprise and question. Why is this community more skeptical?

Re: More importantly: is speed what really matters?

This seems to make 3 assumptions:

First that Safari is the greatest rendering engine in terms of safety, better than anything else that could be deployed on iOS. What makes anyone think so, other than MS acting as an anti-expert of sorts here? I happen to think glitches and JS and CSS standards non-compliance are an indicator of browser immaturity, unless those bits of the standards themselves are security-risky.

Second that I would use Edge. Or Chrome. Most certainly not. Firefox with its own rendering backend is what I’d prefer. Once it had proven robust - see next point.

Third, that I would not privilege safety. I would, but I’d much rather base my choices in that regard on industry experts’ say-so (hint: not MS) than a blanket fiat Apple nyet that ensures they don’t really have competition to challenge them.

Last, iOS itself is pretty robust safety wise. Any _actual_, _demonstrated_, risk to the OS itself should indeed result in a browser’s ban.

Re: Waterfall process much?

> PEP politics and process

Well, PEPs can be fairly anal at several levels. First goal is to keep the syntax clean. There was for example a big old discussion of how to represent C’s ternary operator. Everyone agreed on the idea itself but there was split between “use the C syntax” and “make it more obvious”. The pattern-matching match statement, more or less lifted from Rust, took a long time to gel.

On the tech side, the GIL-ectomy proposal, to remove the Global Interpreter Lock (which holds back thread concurrency) was long in gestation. The concern was: don’t break existing code, including on code that relies on C modules. And don’t screw single thread speed or blow up mem use to enable multithreading.

These things take a long time to get right. Git branching and merging may not be as helpful as you think if the JIT code is embedded finely in critical, and changing, blocks of the main execution engine. Which, IIRC, is “just” a big ol’ C case looking at all the incoming Python operators.

This is mentioned in the article:

> Asked whether development could continue in a fork, Shannon said it was not easy due to the way optimizations are generated, leading to very large code differences that are hard to manage.

Past forays into big structural code changes have been very carefully weighted, and planned, after the execution teams had done lots of initial prrof-of-concept work.

And if you feel you need to diss “waterfall” i.e. planning in this case, I guess that means you’d want “agile”, hmmm?

Re: clarification requested

To be fair the number of websites that can’t be arsed to strip out spaces in a credit card number is even more impressive.

Re: Be careful what you wish for.

adding weird pointless small requirements in public sector often happened from mid level managers or biz analysts, from what I saw:

- higher ups don’t really know or care, as long as the buzzwords are ticked off

- the worker bees are rarely asked anything in depth. If the preceding system isn’t too catastrophic, they’ll say “more of the same”. Generally, they’re like mushrooms and programmers: keep them in the dark and feed them shit. Mostly they want it to just work and minimize change. Many projects flounder because the mgrs and techs are all circle jerking and don’t involve them.

- mid level managers and biz analysts however see it as a unique chance to prove their importance. Out-of-the-box application behaviour always needs “customizing”. Sometimes radical surgery that breaks the app because of some nitwit idea that the sales guy agreed could be done.

Re: Memory

Guys these are ECCeptional circumstances.

For the greats’ good, 6.40GB ought to be enough for the hoi polloi.

Re: What AI did next

... and it will use a lot less makeup in doing so.

Re: Horse

We rely on external libraries to do stuff nowadays. So might as well secure dependencies. Will this work? Won’t speculate, further than some smart folk are trying to lock a door other clevers are doing their best to force. Worth trying to fix however.

If you disapprove so much of 3rd party dependencies, let’s see you trying to roll out your own crypto ;-)

Re: Proud Luddite

The economic and environmental efficiency of driving a $15 burrito in its own taxi to your place so you can pay extra for it are baffling. At minimum wage in most places, the driver’s fees plus Doordash are going to bite hard on such a low base, once past the delivery startup’s loss phase.

As you go up the food value, the economics get somewhat better, but the food deserves better and you could have sat down at what’s perhaps a nice place and been out and about.

Even Snowcrash didn’t quite foresee this:

> "When it gets down to it–we're talking trade balances here–once we've brain-drained all our technology into other countries... There's only four things we do better than anyone else: music. movies. microcode (software). high-speed pizza delivery."

p.s. not in any way to diss on the real poor, but I wonder how many paycheck-to-paychecks are big on using Doordash and making big payments on the oversized SUV they think is cooler/safer to drive to the mall with. Even when it was new - it’s an ‘09 - my Civic is usually one of the humblest vehicles in any parking lot.

Re: Well said

Mostly I remember renting a Ford on 2 different occasions with some truly horrendous MS music/entertainment software (Sync). Basically never managed to play my phone‘s music, despite connecting.

One which got canned by Ford years later, but not before noticeably driving down sales on the cars equipped with it.

Re: Maintainers throwing in the towel?

Kinda agree. A tool that can be used wisely or foolishly.

Currently teaching myself Rust implementing what’s essentially a CLI line editor. The AI writes the code, but I design it, function by function, while reading the Rust Book and watching copious amounts of YouTube Rust info. From time to I drop down into the code and write some. In fact, I’ve started some new features handwriting them on a brand new mini-program in isolation, then asking the AI to merge in my prototype. Design decisions get done separately, chatting in the browser with a different AI.

This is a program I have long thought about writing but recognized Python was unsuited for.

I am aware it’s a bit of “training wheels”. Yet I am being genuinely productive getting started with a brand new language in a month or so and I do mostly understand what’s written because I diff everything it does, revert it whenever it gets stupid and design the architecture. Very little vibe. If anything, on my own, I feel more comfortable with Rust concepts and coding after a month than I do with Javascript after 20x the time coding it.

I also know that this approach is working only because I have written umpteen 100Ks of lines of code by hand before getting here and have researched Rust in advance. It wouldn’t work near as well as a learning tool if I were a noob. I get that, but I enjoy this process immensely. My biggest regret is my printer not working lest I print it out and review the code line by line.

Now I have 2 more ideas I want to run with, both in the programmer utilities domain.

I’ve also levelled up my bash/zsh quite a lot with it as it has facilitated me taking on new syntax which I then write by hand afterwards.

Re: How to Respond

Lucky for Google, the amounts being talked about here seem to be over small claims court limits in many jurisdictions. That’s one get out of jail card, I guess.

</sarcasm>

See, in the French army (late 80s) our field rations included a full pack of Gauloises . Good deal if you don't smoke, can always trade them out for something. I bet a lot of lobbying effort went into keeping them included for as long as possible. Don't remember much about rations proper, my unit generally always had a field kitchen going with it, we had our priorities straight.

Re: Can't wait for this bubble to burst

I am not sure that’s the answer either. Lots of people derive meaning from their jobs. People on welfare don’t always seem to be very appreciative of it. Or enjoy it very much. And that’s before you consider the outcomes with multi-generational welfare.

It might work for some, but is it likely to work well for the majority?

Taken to an extreme you can get a bit of a Morlock-Eloi result out of it. Not sure the Eloi are role models, with or without Morlocks.

Look too at the problems chip makers have in sourcing staff for onshoring chip fabs in the US. Letting skills atrophy in the workforce puts you at long term disadvantages wrt more vigorous countries.

Re: Plank of England

That's one way to read it.

Another is that public sector unions in many countries insist that everyone is on a level playing field and pay isn't determined overmuch by merit. The way to go up is by seniority and/or by management level.

In other words, if you had techies pulling $150K+ salaries to save tax payers hundreds of millions, the unions would likely scream to high heaven. Not least by other techies' unions.

That's not always true. For example, people have long learned that doctors can easily move to more generous locales, so cough up fairly tidy sums (at least past the internship indentured servitude phase). But it is true more often than not.

That's not to say ERP gigs are guaranteed to go right in the private sector. They are not. But at the end of the day, all these businesses, whether private or public, need to learn how to a) source competent staff and/or consultants b) draw up realistic requirements and c) avoid bloodsucker consultancies. The public sector ought to find it easier to coordinate this by rotating staff amongst different projects, since they are after all the same employer while an individual bank will change ERPs every 20 years or so: practice makes perfect. Or at least, not Crapita.

Re: Facebook reportedly acquired or licensed Friendster’s patents

Excellent catch, but that also dates back to 2010 with patents good for 17 year.

Stay tuned…

Re: So are they impounding stuff from tourists on arrival?

When September ends.

Re: Strangely

It was Sybase and when I was getting started on RDBMS in the early 90s, it was far ahead of Oracle (6?) with things like stored procedures and triggers.

Oracle marketing delivered a master class in risk containment however and successfully negated that advantage in a number of shops I saw, as management was always quick to point that Oracle 7 (not yet available and which did not come out till about a year later) did have stored procedures and triggers. Even then early Oracle procedures weren't as flexible as they lacked support for SELECTing data back to the client code.

(whether stored procedures and triggers are necessarily a good idea is another question, to me they're a tool in the toolbox, to be used or not).