On 02/07/12 17:33, Pavel Tkachenko wrote:
Giant mails follow, no panic.
2012/2/6 Gabriel Wicke wicke@wikidev.net:
The enriched HTML DOM we are building (and actually most token stream processing including template expansion) is not tied to any specific syntax or user interface.
It is tied to HTML and it's the same. Even if all of current wikitext features can be represented by HTML (which I doubt) there's no guarantee that this will be true in furutre. This view has probably led to current messy markup.
When I'm saying that HTML can't represent even current wikitext features I imply that we're not talking about microformats and other XHTML tricks. And if we're talking about plain HTML then why not use a completely new format for storing DOM? Or at least clean XML without any standard namespaces that in theory should ease rendering of DOM into HTML (?). It will be parser-specific and won't suffer from future changes of linked namespaces, will be simple to test, etc.
On Future/Parser development diagram HTML DOM is built right after the stream has been parsed into a tree... in other words HTML5 is used to represent wiki. With tricks.
But I'm already venturing into an offtopic discussion here.
But in any case, we first have to implement a solid tokenizer for the current syntax and recreate at least a part of the higher-level functionality (template expansion etc) based on a syntax-independent representation.
I agree on this one.
2012/2/6 Sumana Harihareswara sumanah@wikimedia.org:
Pavel, you're clearly both an intelligent and a technical man - but not all intelligence is of the same, technically-minded type, and it's not always backed up by pertinent and complex knowledge.
I'm flattered with your words, thanks, Oliver.
However, this does not explain why at first Wikipedians had no troubles editing (and even creating) articles and now they are gradually loosing this skill. Is this a result of general degradation?
Not necessarily. It doesn't mean that lose ability to edit, but it can mean that the subset people which know how to edit no longer have available the topics they can write about. (this is just a reformulation of your below argument)
There are two subsets: * People able to edit. * People which can add knowledge.
At the beginning, the interesection was huge even if ability to edit was low, just because there was a lot of knowledge missing. So as the knowledge increases (eg. linearly) "people" appear to be more and more stupid for editing.
I would hate to think this way and believe it's more what Yury has already said above - the project is just getting mature and, naturally, subjects for new articles that are left require more than a general knowledge while edits for existing articles are either complete, require some special knowledge as well or are plain unmotivated enough - new page patrol, "article babysitting", etc. are all "dirty" work and by definition are not that interesting as adding a new article section, prooflink or even correcting a simple typo. (...)
The complexity of our existing markup language is a barrier, but not as much as the presence of any markup language whatsoever as a default.
Now this is something specific to argue about. I must admit that your speech has given me something to think about; perhaps you're right and the fact that initial editors of Wikipedia have come from that "first wave" of the Internet users - with this in mind it's understandable why their number is wearing out.
The usability studies that you have referred to speak with one accord that WYSIWYG is a must. I admit it sounds appropriate in that context. Still, another link suggests that even non-technical people were able to edit and (uh!) format text as bold and italic given a bit of help. And then it notes that even before doing any edits - or seeing an editor's window, be it text or visual - people were confronted with dozens of guideline links and warnings.
Which problem is more important? How you're going to present users with warnings in an inline visual editor? Or is it easier to just put "I've read and understood the rules" fobber-off and consider the matter settled?
Ability to edit and knowledge of rules are probably orthogonal. And users have a inmense rule-blindness. They won't want to read pages and pages of rules or tutorials. They just want things done (eg. change a birth date) I think most people act the same way. What was the last time you read the VCR manual? And we should take that into account, too, not making Rules/Tutorials that look like EULAs. Nonetheless, such thing would be hard to do.
More things to ponder about before my peaceful sleep, huh.
p.s: I wonder why people who can actually give answers are quite often not in the mailing lists.
2012/2/6 Amgine amgine@wikimedians.ca:
As I understand it, for the foreseeable future there will be a raw wiki syntax interface available. I hope contributors can be reassured on this point.
Combined with: 2012/2/6 Trevor Parscal tparscal@wikimedia.org:
Make significant changes to what Wikitext is and can do
The problem with this is that if present "raw wiki syntax" will be kept it will ensure that edits continue downfall.
I disagree. Its existence in the backend shouldn't influcence it.
The concern I see being expressed, fundamentally, is "I have developed skills, practices, and efficiencies with current Wiki syntax. Is your new parser going to destroy my investments in learning? am I going to have to start over with this new system?"
I think it's close in words but not in the meaning. What will you choose: cope with an old dusty car of your grandfather with annual repairs, dyeing, cleaning or find a free day, go to a nearby shop, choose a top-notch car with nano-tech-driven automatic repair, dyeing, cleaner that will serve you for the foreseeable future?
How many programmers (given the opportunity) choose to maintain old spaghetti code over refactoring it to something they'll have pleasure working with?
Quite few, as you probably know. It's the same with common folk who'll stick to old printer, scanner and copier than a new all-in-one device. But it's not right and everyone knows that it's better when they break this trend.
Your proposing is like adopting a new, highly improved C² programming language and throwing all C code (which would be incompatible with the new one). You are not proposing to create a new C² language, but also to stop all support for C.
2012/2/7 Jay Ashworth jra@baylink.com:
Correct, and it isn't merely investments in learning; there are likely investments in wrap-around-the-outside coding which assume access to markup as well. Not All Mediawikiae Are Wikipedia.
I hope this was not a case for keeping old markup running. Most of the time it's better to provide backward compatibility module running on top of the new system than to fix and repair the old system trying to pursue mythical goal of supporting old versions.
Look at C++ STL and what it has become since '89. Look at Microsoft Windows and if its performance on an 4xi7 core has scaled along with Windows 95 on an 80356.
2012/2/7 Mihaly Heder hedermisi@gmail.com:
the millions of pages we already have is not easy to convert in the absence of a formalized wiki grammar
Indeed, but this can be solved by bringing together all pieces of modern wikitext under one roof and building a new strict grammar apart from that. Then a converter can be written that will seamlessly transform old syntax into new and warn user when this is not possible.
So you say "Oh, no problem. I will make this wonderful C2C² converter that will seamlessly produce equivalent C² code from the original C one, so you don't need to rewrite things from scratch. It will be automatically taken care of." Yes. Until that inline assembly that worked in C code makes a random memory overwrite in the kernel. And that other function, which got a compatible by luck of sizeof(int) == sizeof(void*), now in C² makes the application die horribily... and so on.
From what I know this is the direction WMF is going.
and some of them are already afraid that this skill will be obsolete because of the new editor, like the thread starter
This is the second time this argument appears in this thread but I don't understand it. Will you be afraid to "lose" your old coat that has worn our in a bin?
If people is concerned about it, that's a reason to be concerned even if it wasn't rational. I think the reasoning can be explaines as it's better the devil you know...