WIKI WRONGS I woke up a bit early this morning, possibly a bit of routine from 5AM starts on the last couple of weekends in order to help with some burn-offs with the fire brigade. Anyway for some reason I'm in a grump, so it seems the perfect time to write a post that I've had in mind for a very long time, pointing out things I don't like about Wikipedia. In a way Wikipedia is an odd target for me to pick at, because comparing against all the other websites that are widely popular, it's probably the one I like most. I turned anti-Google years ago and now avoid them as much as possible, never used Facebook, Twitter, or any of the other social media websites, and even YouTube has pissed me off so much with their new Javascript-only website that I now rely on youtube-dl as the only interface for it (and mainly just once a month, when I use up my surplus internet data on downloading a bunch of videos at once to watch later). For the sake of buying and selling things, I do use a lot of Ebay and Gumtree, which are also the most common cause of me switching out of the fast and cosy Dillo web browser into generally uncooperative Firefox. Gumtree annoys me as a buyer, particularly with their incredibly crippled search (I think deliberately crippled because they're owned by Ebay), and Ebay annoys me in 100 different ways as a seller, although many items that I sell simply don't in any other place that I advertise them (I wasted lots of time and energy establishing that fact). Yet I feel like all those other popular websites are for a different crowd, they're commercial interests going after the average idiot. Complaining about them not being right for me is like watching a foreign film and complaining that the characters behave strangely. It's just the nature of the thing. However the non-commercial world of Wikipedia always felt like it might be for me. After all, it promised that it might even be made, in part, _by_ me. It's also usable without a modern web browser, although as I complained about in 2023-02-18Dillo_Loses_its_Style.txt, that feature seems to be on the way out now too. However over the years I've steadily accumulated some strong dislikes towards it. In a way I was taught to be something of a Wiki-skeptic. As I grew up in the internet age, a key message from many teachers handing out assignments was not to use Wikipedia. This was in part a valid point about it not being a valid source for reference material, and in part just a generational frustration from teachers who would have preferred we found information from books like they did at school. However after saying this, the teachers were too lazy to really care much about seeing kids browsing Wikipedia, and the only thing that most kids thought about was finding a less obvious place to copy/paste paragraphs from. But being, as usual, an obsessive servant of the local authority figure, I took it to heart and religiously ignored anything except the external links on Wikipedia pages (these days I take a lot from references too, but of course they only make sense if you corrupt your mind by reading the unholy words of the articles themselves). Now outside of school, this practice didn't stick. Wikipedia is just too damn convenient, especially for topics away from computing where few knowledgeable people are willing to set up their own website about them. But fundamentally I still have a distrust of everything written there, and that's prompted me to look behind the references, onto the talk pages, and into deleted articles, and I've often been quite disappointed as a result. Wow I'm thinking too much about what I write here, I thought two hours would be plenty of time to write this and now I'm on 1h 20min and I've barely got to the point. Prepare for writing quality slippage starting now... One example about a general enough topic that I won't be scared about it compromising my anoniminity is the Wikipedia page on PHP. Currently, 2023-03-27, (and it's been like this since at least mid/early last year when I looked into PHP 8) this has the statement (under "PHP 8" - "Just-in-time compilation"): "Additionally, the JIT compiler provides the future potential to move some code from C to PHP, due to the performance improvements for some use cases." Woah, PHP can be faster than C now? You can move code from a compiled language to a scripting langauge for a performance _improvement_! PHP is written in C/C++! What is this witchcraft?! Check the reference... https://kinsta.com/blog/php-8/ The only bit comparing with C is this line: "The benefits of the JIT compiler are roughly (and as already outlined in the RFC): [snip] * The potential to move more code from C to PHP, because PHP will now be sufficiently fast." Hang on, that's not saying PHP will be a performance improvement over C, but that PHP is now no longer so slow that people need to use C code instead because otherwise it would run too slow. Also that's actually just a quote from a PHP developer here: https://externals.io/message/103903#103927 Who was summarising a statement in an RFC proposing that JIT compilation be added to PHP 8 here: https://wiki.php.net/rfc/jit "Lastly - making JIT available can provide us (with additional efforts) with the ability to develop built-in functions in PHP, instead of (or in addition to) C - without suffering the huge performance penalty that would be associated with such a strategy in today's, non-JITted engine. This, in turn, can open the door to faster innovation - and also more secure implementations, that would be less susceptible to memory management, overflows and similar issues associated with C-based development." Therefore the original statement was only a theoretical advantage proposed before the final code was even developed and able to be tested compared to C. So where we've gone from the Wikipedia page is from a claim that some code is now faster written in PHP than C, to some code might not be so slow in PHP that it has to be written in C, to now PHP uses JIT compilation which is theoretically faster therefore if it's all done right in PHP 8 the developers can write internal routines in PHP instead of C and the performance penalty won't be "huge" anymore. So not only is the comparison theoretical, but it's actually a comparison against PHP 7 without JIT support, not against C. It's taken for granted that PHP is slower than C, but with JIT PHP could run fast enough that they can get away with writing less stuff in C internally. What we're looking at here is an explaination of how the developers intend make PHP 8 potentially _slower_ overall by writing internal functions in PHP instead of C. The advantage of JIT is just that they think the performance effect won't be so bad now, so they can get away with it. Also presumably most of the extra PHP-coded internal functions will come later, as this is just the basis for them, so PHP may now keep getting slower over time as more internal functions are added and rewritten in PHP instead of C. So yet another example of how software keeps on getting slower while claiming it's getting faster. Nothing new. But the point that I've now run out of time to make is that this is exactly the sort of misleading trail of half-truths that Wikipedia is full of. Sure, you can find lots of other websites full of convincing bullshit, but Wikipedia encourages the average guy to approximate a bit of something they've read and summarise it without much effective oversight. The annoying thing is that at the same time it discourages editing by people who actually contribute significantly to the project described in the article and can reliably follow the meaning of the source material, because they're considered to have a conflict of interest. Now I'm late to start work so some other points which I probably won't get around to making in a Wiki Wrongs Pt. 2 post: * Wikipedia deletes a lot of pages on the basis of insufficient notability, often permanently throwing away lots of information which could have been extremely useful to small groups of people if it had been hosted elsewhere. See http://deletionpedia.dbatley.com/w/index.php?title=Main_Page (used to be at deletionpedia.org but that seems to be dead, I'm not sure about the status of that version) * Discussion on talk pages seems to be easily overwhelmed by people with the time/energy to just make the changes they prefer and keep re-editing stuff regardless of other opinions. * Pointing out errors on talk pages also often acheives nothing because nobody really cares. The idea of pointing out a problem to be assessed by an expert editor monitoring that page, like a bug report for a software package, often doesn't seem to work. The culture seems to be to make your edit, and if anyone thinks you're wrong, and cares, they'll fix it themselves later. * As a lazy alterntive to self-hosting for many contributors, it potentially saps knowlegdable writers away from building traditional websites, and instead their content is left somewhere where it can later be corrupted by less knowlegdable contributors, or worse have parts snipped away for being too detailed/specialised. * Comparisons, particularly lists, are hamstrung by the requirement (albeit not always followed) to only include items 'notable' enough for their own Wikipedia page, thereby providing an excuse for excluding otherwise valid options (eg. lots of lightweight software that I perfer over bloated junk!) from being presented to someone trying to compare them. * Some software, and related, projects now point to Wikipedia in place of detailed official documentation or their own Wiki. On the basis of the conflict of interest rules this technically excludes involved people from writing their own documentation (although it's probably just that Wikipedia editors don't know who they are). Also it exposes the prime source of reliable information to all the risks outlined above, just for the sake of easy web hosting. Granted my complaint here is mainly against the people running those projects, rather than against Wikipedia. - The Free Thinker