Code in Search of a Home V
Discover more from Rixstep NSA
Oh the irony. For the plan was to time things so part V could begin with a long dissertation on Xfile. This in a series devoted to application domains that can be ported to another Unix.
But if there's one application that doesn't need porting, it's the file manager - except of course for Windows and OS X, the two platforms many of us ended up with.
We have screenshots of the file managers for both Gnome and KDE. Somewhere. In both icon and detail view. Somewhere. Except we can never find them when we need them.
But the basic layout for both is identical. They're essentially made for right-handed people, so the list is on the right and the drop target, the tree, is on the left. Precisely as MSFT Winfile. And as they're Unix applications, they work the same way as our Xfile for OS X (or X-file for Windows).
These Gnome/KDE file managers sport a few redundancies as I remember, they go too far in a few cases, but never extremely so. They essentially stick to the essentials. You need to be able to rename, move, copy files. You want to be able to do this by drag-drop in some cases, I don't suppose they're big on Microsoft's 'copy/cut/paste' paradigm for file ops. You need to be able to set file attributes (modes). You need to be able to generate both hard links and symlinks. And so forth.
Again: you really don't want to overdo it here. You specifically do not want the type of application that makes a John Siracusa or a John Gruber bug-eyed. You want to stick to the basics. And the brunt of your work has to be on improving your code, honing it, making the app more stable and reliable all the time.
Your file manager is your most valuable tool. It must never crash. It must never do anything quirky. Further, it must be known - experienced by you and by everyone else - as not crash-prone or quirky.
Compare this with the latest iteration of 'Apple's finest', where credible people reported online that their system default file manager would crash in such an ignominious way that the entire system had to be cold-booted. How do you ever get in a situation like that? You can't - but they did.
Or take the old Cocoatech forums. Everyone wanted the whiz-bang 'features' of Path Finder. Except they didn't always work, and those forums were full of complaints because things didn't work.
Who releases software like that? If you deliver a product to a bank, or a department of defence, is that what you expect to happen?
Take the annals of Panic, regularly in a panic on a new release. Everything is set, everything is fine, this is going to be a great release! Here we come! Wahoo! Then the bug reports come in. Why weren't those bugs rooted out before delivery? Days and days following that first release - more (unexpected) releases! Long lists of bug fixes! Come on.
Transfer that to an automotive situation. Imagine you just received delivery of your Ferrari F50. It looks absolutely great! Then the first morning after you notice something wrong with the side window. Then with the ignition. Then with the windshield wipers. Is that acceptable? Of course not! So why do developers - both indies and big OS vendors - think it's OK with software?
Beta runs don't exist to ferret out bugs anyway. They're a free form of advertising. Once in a great while they can happen upon something in the bug reports that's of use to them, but generally speaking? No. The testing has to be in-house. PRIOR to release.
Remember when Bill Gates PROUDLY announced the release of NT 4.0 WITH ONLY SEVENTEEN (17) BUGS? Who's kidding who? Perhaps Billy should have a chat with Fred Brooks? He'd never get in the door.
You need test routines and time to ready a release. And, perhaps most of all, you have to eat your own dog food if you can, if circumstances permit (or request). You don't write software in this context unless you intend to use it yourself - otherwise you recruit a sovereign test team who design the test protocols and then carry them out, mercilessly. You leave no stone unturned.
We're currently watching the deterioration in software quality across the board. Most of what we use today, aside from our usual arsenal of apps we create ourselves, are online apps. Specifically, for the sake of argument, we can cite social media software. We currently use four social media platforms, along with several Substack accounts. Guys? Take off your shades; the future's not looking too bright.
It'd be one thing if that software had never been better. Then people would think, like a Jack Handy, 'oh, OK, so that's as good as it gets, OK then, we understand'. As with Apple's NeXT.
But that's not the case. We've written quite a lot about how silly Twitter was. (We've also written about how good Spotify was, so we're not always 'anti'.) Given how toxic Twitter became (and remains, according to what we hear) a number of new platforms have emerged. Most are built on the open source Mastodon.
You can join Mastodon 'as is' (there's a thought) but you can also yank the source and build your own. Something three of the four platforms we've tried have done.
You're going to find this surprising, but the most stable, what we've seen, is... GAB. Now a lot has changed at GAB of late - they can't resist adding bells and whistles - but the little we use it, it seems OK. And it's based on Mastodon.
Next is GETTR. Who knows what they're doing there. We were approached by others when first we joined. People told us to watch out for all the bugs. Oh really? Yes really.
This is a recurring theme in software development today. If it's down to not being able to find or afford qualified help, that's not known. But it does look like their programmers are not full-time hires but only called in occasionally. Otherwise it's hard to explain how code changes are rarely tested properly, and why it can take a fortnight to see the simplest bug fixed, despite things being quiet online.
The final site is TRUTH. This is a project run by Devin Nunes, who shared a legal adviser with a colleague. Devin took it slowly. Agonisingly so. First for iPhone only. Then for Android too. But always only in North America. Then it spread to other platforms and eventually to all countries. So one expected things to be very very solid.
And for the most part they are. The ergonomics are brilliant. Generous spaces where GETTR crowds everything together. A few new twists in Mastodon, and they're nice. But the same thing when it comes to bug fixes. Bugs do appear, and they'll probably be around for a fortnight until they can afford to bring the programmer back.
One bug concerned the report panel. The panel rectangle has to be set to expand as the panel's contents grows. Somehow the programmer forgot to tick the box for that. Wait a few weeks and it's taken care of.
Right now they have some kind of synch issue with reposts and likes. Reposting had to fetch the previous count and increment it. As of ten days now, it doesn't get there in time. Somehow the code also updates the like status as well, so if the code fails, then the like is gone too. But something's in motion there, so if you repost again, this time it'll work, but the repost count will be too high! It's been like that for about ten days now. We've seen people complaining. We've seen repost counts go down on accounts we frequent. They too must bring in a programmer every second week or so.
And for what it's worth: Truth would do well to steal from Minds in one regard: limit the number of tags in a post. That sounds pointless but it's not. Maniacs and other purveyors of the unsolicited will pack posts to the brim. Minds caps things at 5 per post.
Another angle is memory consumption, both in RAM (VM) and on disk. Minds is the worst offender here, caching obscene amounts of data. Normally a hard page refresh can be necessary so Minds can work properly, especially when trying to retrieve site data. 4-6 times as much in cache as the other sites. Perhaps it's partly down to a memory leak? What we know, Minds isn't based on Mastodon.
As for Substack? Yes it does a lot of cool things, but it too is deteriorating. Must be the climate around the bay area. No one understands design anymore. Contradictions and conflicts abound.
Anyway, back to code and homes. The overriding point with open source is that if you don't like the code, you can change it. The social media sites are taking their base from open source. Open source can mean bad things, but it can mean good things too.
Next time we'll resume with the story of Brent C Ethington who worked out ot Anaheim California.
PS. Your IVM should be good even for common sniffles (3mg).
https://flccc.net/covid-is-on-the-rise-again-heres-why-you-shouldnt-panic/
Postscript: They can’t do this
Just a reminder. No other file manager on Apple’s OS X platform can do this. For that matter, Steve Jobs’ platform in Redwood City couldn’t do it either.
But Xfile and the open source Linux file managers should be able to carry it off. And you know why, right? Until next time, be good and stay safe!
Afterword
You might know someone who’s interested, or you might be interested yourself.
Radsoft/Rixstep are having a fire-sale on all available products.
Normally the prices are much higher.
Currently they’re going for $72 each.
Why? As we ready a transition to use of open source systems, or at least contemplate it, we’re thinking we might ‘sit this one out’. Too much head-banging for far too long.
Note: You can pick up both the XTP and the ACP together for $144!
Be good!