User talk:Abd

From Beyond politics
Jump to: navigation, search

Where's your pretty URLs? In my experience, the easiest way to fix that is to create an .htaccess file with this content:

RewriteEngine on
RewriteRule ^wiki/(.*)$ /w/index.php?title=$1 [PT,L,QSA]
RewriteRule ^wiki/*$ /w/index.php [L,QSA]
RewriteRule ^/*$ /w/index.php [L,QSA]

The wiki should be in a /w/ subdirectory of the directory with the .htaccess file. Yevgeny (talk) 14:34, 29 July 2014 (UTC)

Okay, the hosting is shared, i.e., beyondpolitics.org is an add-on domain. So there is a directory /home/username/public_html/beyondpolitics.org and mediawiki lived in that directory. Following the above, I moved the entire contents of the beyondpolitics directory to the subdirectory /w/, and added the .htaccess file with the provided content to the beyondpolitics directory.
Naturally, this broke the wiki. So I edited LocalSettings.php to show:
$wgScriptPath = "/w";
And it started working again.
But I still don't see pretty URLs. Could there be a server issue? --Abd (talk) 00:54, 30 July 2014 (UTC)
Actually, I think I know. Brain fault. --Abd (talk) 00:56, 30 July 2014 (UTC)
Okay, it works, for example, http://beyondpolitics.org/wiki/User:Abd now loads my user page.
Oh yeah, I forgot to tell you about $wgScriptPath, $wgArticlePath, etc. Yevgeny (talk) 04:41, 30 July 2014 (UTC)
  • Indeed.
  • ArticlePath is not in LocalSettings, and it's the "etc." that I worry about.
  • Mediawiki documentation is a train wreck, way too often.
  • Sinclair adopted Liquid Threads, apparently not realizing that removing it would break all the discussions started under it. It looks like LT doesn't export.
  • What I've called "consensus reports" are missing. So someone looking up LT to see what's so about it finds a massive mess, discussions and archived discussions, and very little clue. Yes, deprecated, no longer being supported, that's a clue, but the reason why? After all, Why not try it?
  • Consensus reports are generally missing from Wikipedia. A consensus report is not a report of action, though it will include that. It would be, properly, something that precedes action or that reviews action, generating advice. Consensus reports do not make choices, they provide the information on which a decision-maker may base choices.
  • It's work to generate a consensus report. However, when this is done, genuine consensus will be maximized as to advice, and resulting actions, if undertaken considering the report, will unite the community. A consensus report, by design, will be far easier to read than the train wrecks that it might be compiled from.
  • Offiwiki 1.0 was, and remains, a clear demonstration of Wikipedia dysfunction, in many ways. This should not be surprising, at all! After all, he invited the same community.
  • Sinclair ultimately decided to follow, not what he had declared he would follow -- he blatantly disregarded his own repeated promises --, not the community (which gave him readily discernable nearly-unanimous advice), but the advice of some of those who are at the core of original Wikipedia dysfunction.
  • He still has not restored the Move command to ordinary users, while asking for help reorganizing.
  • His central purpose seems to be making Offwiki "safe." I.e., no disputes. No "discussion of users." Yet he has users who discuss users, including himself, and who do so with high incivility and blame. But he "likes" those users.
  • It's a brilliant demonstration, a really good occasion for re-activating beyondpolitics.org --Abd (talk) 13:14, 30 July 2014 (UTC)

The reason I say "etc." is that I have to play around with those settings too when I'm trying to get my wikis to work; I don't know them by heart, there seem to be some that you would only use in special situations, and the pretty URL documentation page has been a mess for years. Also, it's not just MediaWiki documentation that's sometimes a train wreck, but MediaWiki design as well. I guess the "etc." could also mean "and $wgUsePathInfo". Yevgeny (talk) 13:27, 30 July 2014 (UTC)

Right. Mediawiki itself demonstrates crowd-sourcing function and dysfunction. It is a more coherent community, apparently. Or not. I'd bet that, looked at more closely, one would see the fault lines. It is mostly a smaller community. --Abd (talk) 13:50, 30 July 2014 (UTC)
As Brian Wolff wrote, "Submitting a (useful) patch is hard work. This cuts down on the number of trolls significantly." However, like Wikipedia, they can be kinda resistant to change (as Brian Wolff also wrote, "change is scary!") Also, code review can be a bottleneck because everything has to be approved by these guys. The great advantage MediaWiki has over Wikipedia is that it's more modular so, e.g., I can create an extension without having to get central approval.
If you want to get all the extensions, you can try git clone https://gerrit.wikimedia.org/r/p/mediawiki/extensions.git from the command line. Do you have shell access on your shared hosting? Most shared hosting services let you get it, although sometimes they make you jump through a few hoops. E.g., I think one webhost wanted to call me on the phone first, as another identity verification, before granting the access.
I'd say creating a logo should be a high priority, since for one thing it makes it more evident to people which wiki they're on, and it shows you've put at least minimal work into the aesthetics of your wiki. You could browse the Wikimedia Commons for a placeholder logo, and then shrink it down to 135x135. Sometimes it won't be a perfect square, so then you can either crop it, or shrink it down to 135xsomething or somethingx135. Actually, it works out pretty well sometimes if it's wider than it is long, because then you can shrink it down to, e.g., 135x90 and put a caption underneath. Consider dropping shadow behind your logo; it can make it look a lot better.
It usually takes me a few hours to design a logo, because I have to figure out what I'm doing with GIMP. Logo design is a pain in the neck, but pretty much mandatory if you want to look like a serious wiki. Alternatively, you could try an easier-to-use tool like Paint. Yevgeny (talk) 14:12, 30 July 2014 (UTC)

Shell access

Yes, when I bought this new hosting account, I checked that SSH was available. So how?

Way cool. I must authorize it from my Account Manager, by entering an IP address. It will automatically enter my current IP, and only allow access, then for my main user, with that password, and from that IP. I went through an entire system crash, etc., so I needed to download PuTTy, did so, and just transferred the entire extension package with the command you provided. Thanks. Worked like a charm.

Cool. Did you have previous experience with Linux/Unix/etc.? Yevgeny (talk) 13:22, 31 July 2014 (UTC)
Yes, Unix, way back, then I've used SSH with Linux previously as well. Used to edit with vi over telnet. Etc. Used to write programs to do what I wanted all the time, but almost always with QuickBasic. Loved Forth, though. I never developed any web programming experience, beyond using HTML weakly.
I accessed the WELL originally with a 300 baud modem using a Commodore Vic20 computer. I remember spending the better part of $200 or so to get a 1200 baud modem, and used a Commodore 128. --Abd (talk) 13:44, 31 July 2014 (UTC)
So what would that be in today's dollars? I used to have a Vic20, but couldn't get it to work; it was missing the cables. We went to Radio Shack, but of course they didn't have them. My dad complained that the problem with Radio Shack is that they only sell their own stuff (e.g. Tandy products). Yevgeny (talk) 13:48, 31 July 2014 (UTC)
The Vic20 was a cool machine. You could still program it in machine language, you could still understand everything happening in there, and I did. Obviously, you needed cables! They would have been available on-line. I eventually sold that stuff on eBay. The cables, however, were standard commercial products, if you knew what to look for. You might have needed to do some soldering. Radio Shack might or might not have had what you needed, but it would not have said "Vic-20" on it! I wrote my first computer article, however, on the Altair 8800, I built one, from the original Popular Electronics offer, and created "The Impossible Dream Cassette Interface." It required a simple cassette recorder, I used one from Radio Shack, in fact, plus a capacitor, a diode, some wire, and a very short machine language bootstrap loader. It worked, very well. I invented a coding scheme for it. Much later, I used the same scheme for a fiber optic network. It was a bad idea, when taken up to much higher data rates, because the transmission time varied with the data, and it was extremely difficult to debug the electronics without a storage oscilloscope, which I didn't have access to at the time. It was great for a boot loader.
https://archive.org/details/byte-magazine-1977-02 page 82. --Abd (talk) 14:23, 31 July 2014 (UTC)

Bleeding edge

If you really want to be on the bleeding edge, you can download the latest development version of MediaWiki. I've only encountered minor bugs in these development versions. Usually the benefits of having the latest features outweigh the costs. Yevgeny (talk) 20:27, 30 July 2014 (UTC)

I don't want to be "bleeding edge." My blood supply is limited. It's bad enough with the "stable versions," the documentation is so poor. Now, if there were some feature that I saw as a necessity, that was certain to become part of the latest stable release, it's still possible. --Abd (talk) 12:15, 31 July 2014 (UTC)
With a well-designed system, the documentation doesn't always need to be all that great, because the software can easily be figured out without referring to it. The configuration stuff you've been working on so far is the same stuff thousands of other independent wiki owners have had to deal with. However, I think they figure thousands isn't enough to justify a WordPress-like GUI-based configuration system. There's mw:Extension:Configure, but it no work too good. Last I heard, proposals for a configuration database were in their second round. Also, there had been a Google Summer of Code project to create "an awesome extension management platform for MediaWiki, facilitating the installation, updating, removal and configuration of extensions." I notice you did describe your recent experience downloading the extensions from Git as "Way cool" but I don't think that installation method is what Brion had in mind. Yevgeny (talk) 13:24, 31 July 2014 (UTC)
As a programmer, you would think that! "Easily figured out" assumes a lot of prior experience. Sure, I can "easily figure out" software by trying stuff. However, it can take days, sometimes. I will have ideas in my head that are, quite simply, different with those with different experience. There will be an instruction in a setup, for example, that assumes you know how to do a particular task. For example, suppose I didn't know what "git" was. An instruction to "git" some page will be meaningless, I'll have to look it up. If I'm lucky, someone explains it near the top of a page somewhere that Google picks out of the vast levels of noise.
Wikis were invented to develop system documentation. It worked, to a degree. I have not yet installed suggested extensions, because I want to review them before installing, and that could be "fun." Some I know I want, already, but it's the others ...
If Mediawiki were commercial software, intended for general usage, it would fail, it would fall on its face. It's "free," but the cost to the user is high enough, in terms of training, such that it would be more sensible to buy a product, unless one has plenty of spare time that isn't valued much. And wikis are mostly created out of "spare time." If a wiki was for business, and one wanted to use MediaWiki, because of the available user base, one would hire an expert. Which is, of course, expensive.
Could it be that the software developer community doesn't want to create clear, easy-to-follow documentation because this could reduce demand for their skills? Tech writing is, in fact, a distinct and different skill than programming. One would also experiment with totally unskilled users, seeing what mistakes they make, and then designing the documentation to prevent most of them. Hypertext, Yevgeny, don't leave home without it. Wikis could be great for this, still. But it takes work. And, my guess, the programmers dominate Mediawiki. A good tech writer might not be welcome.... The tech writer, by the way, will make lots of mistakes. So a good final product will be a consensus document, the programmers would need to cooperate and collaborate with it. I suspect that they might simply reject an "outsider." --Abd (talk) 14:05, 31 July 2014 (UTC)
Oh, I didn't mean to come off as saying that you should find configuring MediaWiki easy to figure out. If you ever saw my first wiki, you know that there was a learning curve. Git is also pretty tricky to figure out; I only know a few things about it, and when I use it, I probably make the process way harder than it needs to be.
By the way, did you know you can run queries on Wikipedia's database, or a replica thereof? They talk about it here. I think if you run the right queries, you can find out stuff like how often sysops block themselves, etc. Once you have access, getting to the place where you do queries takes about three steps, but I think it takes me about a half hour to figure it out each time because I keep thinking, "Ah, no need to write this down, I'll remember it." Yevgeny (talk) 14:13, 31 July 2014 (UTC)
Yes, I know. About all of what you wrote (except the specific Bastion access link), including failing to write it down. I still think I'm a teenager. I'm not. In fact, I stopped being one fifty years ago. I remember when I became unable to hold two major threads of consciousness at once. It was over twenty years ago. I.e., I could listen and talk on the phone and follow the other person, fully, and read a book at the same time. And then, out of the blue, it seemed, I had no idea what the other person had been talking about, and I'd probably been making "communication noises," i.e., "uh-huh," etc. It was totally embarrassing. Doctor: normal aging, get over it. --Abd (talk) 14:36, 31 July 2014 (UTC)

Paths

Try:

$wgScriptPath = "/w";
$wgArticlePath = '/wiki/$1';
$wgUsePathInfo = true;        # Enable use of pretty URLs

Yevgeny (talk) 16:20, 5 November 2014 (UTC)

Try these two extensions

Like ParserFunctions, Scribunto is pretty much a must-have if you want to be able to import templates from Wikipedia and have them work. Yevgeny (talk) 23:26, 4 April 2015 (UTC)

robots.txt

I forgot to mention, you'll want to create a robots.txt file that says the following:

User-agent: *
Disallow: /w/

Yevgeny (talk) 23:30, 4 April 2015 (UTC)