Coffee overdose

For goodness' sake, it's half past four in the morning.
Been a while since I'd written anything.
Long story short — the usual issues are still there, exacerbated by the whole COVID craze rather than quenched by a year's time spent being holed up at home.
Going back to the office, albeit only for three days a week. Now that's some fresh air at long last. I must admit, I friggin' hate WFH.

I wonder what is the actual matter with the binary parsing.

I mean, it's not that much more particularly hard than the bona fide text parsing. It does get a bit convoluted when error recovery is involved, yes, but in the end, why not? You google up text parsing and you've got yourself a whole lot of stuff, ranging from small parser combinator libraries to full-blown parser generators with bells and whistles like ANTLR. But try the same with binary parsing and it's all half-assed solutions which can't do bit twiddling or are written by someone trying to go all declarative (unfortunately, json kind of declarative, fields and structs) without going into any of the uglier binary format issues.

 In the light of the previous post, I've been digging around all the budgeting apps, and all of them seem to be grossly lacking functionality in the budget department, i.e. go find an app which can do low/high income projections, or tie your budgeting periods to paychecks, or tie the envelopes to percentage of income instead of fixed sums, or having a shared account, or letting you play around with projections like savings/mortgage/etc.

Kinda stupid. All I see around are glorified expense trackers. YNAB kind of hits the right spot for a few things, but mostly it has the same problems, along with being on a monthly subscription fee which is not exactly a thing you want when you're running on fumes.

I'm kinda thinking of my own app for that matter. Expense tracking with shared account and basic envelope budgeting would be a start, I guess.

Saw some movement in the news about Tim Berner-Lee's Solid.
TL;DR: Hard to figure the whole point of it at a first glance.

---

As a matter of habit, I'm inclined to feel good about it, but let's be honest it looks like another piece of turn-everything-into-RDF-semantic-web-tinfoil-kill-google-with-fire-now kind of bonanza. Wouldn't exactly be the first.

On the other hand, the premise of hosting all of your data on your location of choice seems rather legit. The one thing I don't understand though is how it is going to be enforced in practice. Open source, data retention audits, yes, good, more of that please. But still, what's to stop another corp from forcing a user to provide certain data in exchange for using their services? From collecting it in the process? No bloody foggiest idea how turning everything into an RDF galore is going to fix that sort of mess, which is, seemingly, the main reason the project even exists in the first place.

I guess figuring out what they are trying to do is going to take a little more digging around the documentation. Aw hell.

The WiFi chip story of a few recent days concluded with some ugly soldering. But it didn't end there, nope! The chip went up, but the stock driver crashes. Enduring craftsmanship, so to speak.

Well, not that there's anything else I can do about it at this point, so, at long last, I can put the issue to rest, at least for a while.

 I'm wondering if I'd be able to get the damn MediaTek driver to work on the target hardware, given all its dynamic DMA buffer requirements. So far even decreasing buffer size while increasing CMA size didn't prevent the allocator from failing. Sometimes embedded development can be a tad too much fun, to be honest. But hey, I'm not even making MCU firmware here, so I guess I shouldn't complain — after all, you could run into all the same issues on a fully blown PC around ten years ago. :)

That funny moment when the morning is so lazy you need a long stick to fetch some from the bookshelf without getting out of bed.

The LJ is down once again. I keep wondering what the hell is going down there. The poor thing has been DDoS'ed much too often recently. It seems that it's more convenient to stay here on dreamwidth and keep exporting to LJ.

Однако. ЖэЖэ в последнее время стал уныл и нестабилен чуть более, чем полностью. Перебираюсь сюда.