11 March 2026
Why do 'developers' have to keep sticking their oar in on things that are working perfectly well? Currently on the Linux desktop I am struggling with things like the main menu being missing from the LibreOffice window, appearing instead only on the start bar if I am using the middle screen. The same on some other apps such as Dolphin, while other application are unaffected. The problem with those applications is the failure of the cursor scaling, so while hovering over the Cura working area it is almost impossible to see the cursor, which has been scaled to display on my display properly on my setup. This consists of 4 4k monitors, although only three are actually on the desk, with the fourth channel feeding the TV set. Having now sorted that to mirror the right hand monitor, at least most of the time it's not a problem, but switching the TV on and off causes the main displays to blank a few times before the computer decides what the 'new' setup is. That particular 'helpful feature' was a major headache when I was building display systems, where the control monitor was not actually plugged in, but was accessed via VNC. Thankfully dummy monitor dongles came to the rescue, and I'm actually using one of the HDMI ones today to get the current setup even working. Then we move on to Wayland over X11, and the additional problems that is causing since some applications have not yet been 'ported over'. I'm still on an X11 session at the moment and being told that all my problems will be solved if I use Wayland, but that is not what I have been finding, although perhaps if I wiped the machine and only installed Wayland things may be different, switching between the two seems to be at least some of the difficulty.
The next problem that is slowly creeping up is the retiring of YAST. I get all the arguments but neither of the servers HAS a desktop, so using the better 'graphic tools' is not an option and currently I can manage everything via a command prompt window with SSH to each machine. Yes we can drop back to command line tools, but YAST ones just work so well ...
I started to use PHP at the end of the 1990's when my main client, the Benefits Agency, started to use PC's in the offices. Up until then, my single PC with multiple remote terminals was often the only computer in the office, with everything else being handled manually by looking up cards in the card index. Our Caller Management System used anonymous ticket numbers to keep track of clients in the office, and waiting times were measured often in hours! The move to having an actual computer system allowed the replacement of the remote terminals with a browser page, and PHP was a lot more flexible than the compiled BCB code that preceded it. That was when I started using the ADOdb abstraction layer to access what by then was Firebird, which had replaced Interbase after Borland decided that that should be end of lifed. Once they realised just how critical Interbase was to major systems around the world, they tried to close the door, but the open source version is well supported today. The data that I stored back in the 1990's is still accessible today! It's a pity that the continual drive to 'improve' PHP and destroy some of it's inherent simplicity is a problem today. My current setup is still suffering from areas of the code that worked fine 20+ years ago, but are no longer 'politically correct'. Having to flag every variable that is operationally perfectly acceptable to not have been set is an ongoing minefield, and no I will not hide those 'warnings' as from day one my codebase had always been clean, so any hickup was important to know about.
The path to achieve what I was planning back in the last century has still not been achieved, and there is nothing 'off the shelf' that answers all the functionality that we were looking for back in 'TikiWiki' days. A modular framework to which modules could be added as required lead to the split that became Tikipro, which was later renamed Bitweaver. This formed the base of my commercial software for many years, although the Caller Management System never ACTUALLY used it live as those clients were happy with the older system. Various pressures lead to the end of it's use across over 250 offices at one stage, starting with the sale of those offices to private caretakers at which time while the IT system remained with the Benefit agency, the Caller system was 'part of the office' and the new owners simply stopped paying for support. That said, many offices did pay for it out of their own budget, and the next hickup was the death of XP when Microsoft used it's illegal monopoly to force many government departments to 'save money' by agreeing to take on the new licencing structure and their 'client management system' which never did provide the call system that was promised as part of hte package. Again some clients carried on funding the CMS, but we were down to 10 or so sites, and the last of those, Wyre Forest', only retired the system after Covid, when they they closed the drop in centre completely. Although the phone system had been using it, that had at least evolved the tracking tools that replaced the CMS data.
So where are we today? I have retired and am now only keeping a few of the legacy websites alive, but it's a continual battle against PHP and the vast majority of packages that were available for bitweaver have not been brough up to 'modern standards'. I am fighting to keep those that I do use active, but have had to give in and Nextcloud is running to provide a stable interface to my mobile devices and I've got Webtrees working with Firebird so I can access data from other bitweaver packages, but the legacy email archive is still only accessible via thunderbird on the desktop machine and even Nextcloud has no facility to access it. I need to review just what work is needed next to try and plug the holes which have been open since day one of moving to web based services!
