Follow

my most galaxybrained small thing:huge effect opinion is that the Wikipedia Vector skin was a catastrophe for digital culture. Wikipedia, always an encyclopedia first and wiki second, introduced and innoculated a generation of people to wikis in a way that completely de-emphasized all the radical parts of wikis to appear like an encyclopedia. NO ONE KNOWS about "what links here," "wanted pages," or talk pages. so both the graph structure and dialogic reality of wikis is LOST. !/

taken in context with the historical moment, nupedia-> wikipedia's rise in the early 2000s, as a time when people still mostly interacted with individual websites, people were introduced to Wikipedia, not meatball wiki. so as geocities morphed into myspace into Facebook, there was no alternative (but forums) to creating your own digital space. Because people did not see wikis as the infinitely malleable cultural space they are, we got social media

the practice of wikilinking is transformative if you experience it like a wiki head. as you write, you [[wikilink]] your way through [[basic concepts]] that you know will be [[wanted pages]] at some point, so in the act of communicating you are also building the structure of information around you. by properly representing inlinks and outlinks, you completely explode the problem of "where does this information go:" because you will always be able to find it one hop out

(that is not a [[sorry anagora]] moment because those were totally on purpose).
And that, combined with always mutable page histories, is an extremely liberating thing!!! you literally never need to wonder "[[does this go here]]," "[[how do I write this]]" because the answer is always [[Yes]].

but that, and the fluid creativity of discussion happening amidst the "factual information" is completely lost with Vector!!! it just barely exists. If we had something like Obsidian in 2002, the internet would be different.

@jonny I am unclear about this discussion. The way I see vector is that raw editing of infobox, templates are extremely dated concepts when every other editor(except competent MD/LaTeX) editors have adopted WYSIWYG. Even .MD is moving in WYSIWYG mode and latex is used much of the time for a single line of maths.
Vector skin feels a lot cleaner to me than it's predecessor, you can go back to the classic of course(I understand the power of defaults).

@jonny

An opposite approach was something that I first saw in an app called Tomboy, a decade ago. There was no link syntax. Any existing note with the same title was automatically linked (some wikis call this "radiolinks"). Any text could be selected and turned into a link to new note. There were no red-links/wanted pages, because every single word could become a new page easily. Every page was wanted. It was probably the best writing experience I've ever had. Org-mode or howm doesn't come close.

@Sandra
interesting, this sounds like everything2, where each page could have multiple individual versions of it. how were delimiters between words/phrases done? and what was the notion of "page" titled under? or I may be misunderstanding

@jonny You selected text to start a new note with that text as the title (there was also a keyboard shortcut to start new notes and just type in a title) and then all occurances of text that was the title of another note became links to that note. Pretty great ♥

@Sandra
whoa and overlaps??? like if you had
multiple links to pages
And [[multiple links]], [[links to pages]] and [[multiple links to pages]] it overlaid all 3??

@Sandra
and wait was that then reflected in syntax or was it a wysiwyg???

@jonny It was classic a 90s style wysiwyg UI design. But I wanna make a similar radiolinks setup in Emacs or something.
@jonny

I don't remember, but what would make sense to me is to prefer the long one.

Like, if there's a mention of bad movies, you'd want to go to the note named bad movies and then from there go to a note named just movies.

@Sandra
Cool! Sounds like such a UI can be the front-end to a wiki.

@jonny any thoughts of federated wiki, also by Ward Cunningham et al?

@jonny Also, that every wiki has to be basically Mediawiki. I've had this discussion about Emacs Wiki a few times.
alexschroeder.ch/wiki/2012-03-

@kensanata
yes definitely. I think the wiki community from the very start has always struggled with deployment and implementation. that many contemporary program languages are strongly reorienting around packaging and dependency management, npm is so influential, etc. is because it has been the missing development stage for most technologies. like wards wiki was some totally custom script that only ran on his machine right?

@kensanata
that they are reorienting...is no accident**
like mediawiki, even though it's still unnecessarily hard to install and configure, is still extremely simple to deploy: download some sql files, have an SQL daemon running, and done from any PC

@jonny It's true. Also the reason why my latest wiki is packaged for the language (in my case this being CPAN for Perl).

@kensanata @jonny

Latest corporate Wiki irritation...

Good news - I actually manage some enlightened techies who are keen to share knowledge, save time and improve efficiency for the team and, potentially, a wider audience.

Bad news - they keep sending me 'HOWTO's in Word format.

'But I don't like Confluence, it's much quicker for me to write a Word document'.

'Well guess what. I don't particularly like Confluence and believe $CORP should enforce Emacs and Orgmode on every single employee but guess what, this isn't gonna happen. Thanks for the useful HOWTO. I'll struggle to extract the images and fix the formatting for Confluence later'.

@andyc
@kensanata
like I completely agree that is infuriating and have experienced it many times myself, but it does beg the question why it is easier for them. inertia yes, but the "edit but I can't see it until I build/save it" model is scary for a lot of people. files are easier to reason about than wiki pages because they are "there," where wikis have relied on being self-organizing with page, rather than navigation as the prime UX element, etc.

@andyc @kensanata
I have had luck with integrating with process: writing in word and then converting with pandoc is an easier switch for some people. with Mac I have a lil automator action thing so you can right click and convert to wiki and then just copy and paste. even being able to work off a cloned filesystem is easier bc they can edit files locally and sync rather than use browser tools which often feel flimsy.

@andyc @kensanata
like I definitely hit edit, copy and paste the whole source into a text editor, edit, and copy and paste. column selection and all my regex stuff and syntax highlighting and indenting and completions are there yano.

@andyc @kensanata
I really love obsidian lately. and if it had more of a hostable wikilike form I think it would basically be a very powerful way of approaching the problem. particularly if it finally made transclusion as fluid as a filesystem metaphor.

@jonny @kensanata same reason most folk are more comfortable in excel than a database. It's easier, familiar and they spend a lot of time using that software.

Still I'm just grateful they're creating content. Baby steps :-)

@andyc @kensanata @jonny we use confluence at work and while I'm not a big fan the stuff we keep in it is a lot more useful than our old (moinmoin!) wiki where all sorts of stale stuff went to die. For individual project landing pages with easy jumping off into the jira they work well enough.

@stsquad
@andyc @kensanata
what makes it fresh? like is it the use with jira that does that? or different patterns? not that familiar with it

@jonny @andyc @kensanata integrations generally, for example: linaro.atlassian.net/wiki/spac easily shows calendar info and latest JIRA tickets. I've managed to hack similar on #mediawiki: wiki.qemu.org/Testing/CI but it's a lot less flexible.

@kensanata @jonny I always find it strange when people moan about 'Wikis' - because the entire point is that they are supposed to be *collaborative* writing efforts.

I suspect you have probably had less than 20 contributors to EmacsWiki and that you don't sanitise or mindlessly delete content.

I often see this mentality in the corporate environment too.

'The solutions invariably have nothing to do with the problem.'

@kensanata
what wiki software interests you? like what should I check out that does things differently? I have mostly read wiki code from a historical perspective

Wikis 

@jonny I develop Oddmuse where it is important to me to be very hackable via extra code in config files and "modules". Sadly all outside of CPAN, but hackability has allowed me to do:
emacswiki.org and oddmuse.org (classic ephemeral half-organised documentation)
alexschroeder.ch (blog)
campaignwiki.org (namespaces: each game gets its own wiki)
communitywiki.org (a bit more forum like with the discussions and the little avatars)
But also special purpose sites I wanted to experiment with (didn't go anywhere):
communitywiki.org/food (recipes)
alexschroeder.ch/software (bugtracking)
alexschroeder.ch/paper (publishing)

Phoebe started out as a Gemini wiki which offers the same kind of hackability but all of it well packaged with Perl modules, and using a web framework that includes its own web server.

Wikis 

@kensanata
thank you!!! I have to take a look, I don't read the language but I am gonna try

Wikis 

@jonny In that case:
alexschroeder.ch/cgit/oddmuse/
alexschroeder.ch/cgit/phoebe/

No longer of interest

This one was written in Raku back when it was still called Perl 6 and I was thinking about the future:
alexschroeder.ch/cgit/oddmuse6

This one was written in Go so I could learn the language:
alexschroeder.ch/cgit/oddmu/

Both were abandoned because I could not see an easy way to add local markup rules in a config file that would end up being well integrated in the rest of the parser.

@jonny But that is the normal process of denerdyfying any cool thing out there, no? To become popular, it first has to become shallow and boring.

@deshipu
there is some amount of need to get it out of the hands of zealots and into the hands of many! but I don't think it necessarily always needs to lose its core features to do so. I think this one was a specific feature of the cultural evolutiom of wikis through an encyclopedia, which they are very good at, but the encyclopedia they created obscured the culture it was born from

@jonny Personally, I was always on the "wikipedia is not a wiki" side of that culture, so to me it's just an accidental remnant of a movement that died independently of it, killed by the SEO phenomenon. Since radical openness was one of the features wikipedia rejected, it was somewhat immune to that, and so it survives.

@jonny
I've been editing Wikipedia since over a decade and I don't know what you're talking about :(. I don't even know what a vector skin is..
Don't worry, I don't need you to explain, I could easily search the web if I wanted to. I'm just saying that even relatively experienced wikipedians don't know about whatever awesome things you're describing. Quite sad.

@maltimore
it's the default way Wikipedia looks ❤️❤️ (now with Minerva as the mobile skin). idk if they have any other ones enabled, but on most mediawiki instances you can go to your user settings and select a skin. see tweeki as another example of a different era of web design (not exactly the best) applied to mediawiki

@jonny It's a Wikipedia that anyone can edit so it's not really to be trusted anyways

@whitepaperkat
ok hi you made me laugh a lot and then more nice 2 meet u

Sign in to participate in the conversation
social.coop

A Fediverse instance for people interested in cooperative and collective projects.