The Visual Editor is a huge failure

[Randy from Boise]

Fram doesn't pull punches. For that, I respect him.

RfB

[Vigilant]

Fram doesn't pull punches. For that, I respect him.

RfB

It's even better that he's incontrovertibly correct in these assessments.

[Kiefer.Wolfowitz]

Fram doesn't pull punches. For that, I respect him.

RfB

It's even better that he's incontrovertibly correct in these assessments.

Agreed. Power, Randy. Precision, Vigilant.

Also, a quick return to his basic position to avoid having his arm trapped.

[HRIP7]

Balls.
https://meta.wikimedia.org/w/index.php? ... id=9972892

Ballsy indeed. And cogent. (No reply to date.)

[Vigilant]

Unexpected behaviour editing a link

Suppose I want to insert a link to Foobar but mistype it as Barfoo. I spot the error and edit the text, which is now a red Barfoo, so that it says Foobar. What I get is now [[Barfoo|Foobar]] which is certainly not what I intended. This seems counter-inituitive. Deltahedron (talk) 08:19, 27 September 2014 (UTC)

Totally not possible to catch that bug in a regression test!!!!!!!

Can the WMF engineering department do anything right or is this their collective hell that we've stumbled, unwittingly, into where they are destined to repeat forever, ala Prometheus, the worst development cycle in the history of the universe?

The mind boggles at what fiendish act they must have committed on the mortal plane to deserve such heinous censure.

[Poetlister]

The mind boggles at what fiendish act they must have committed on the mortal plane to deserve such heinous censure.

Work for WMF?

[Vigilant]

The mind boggles at what fiendish act they must have committed on the mortal plane to deserve such heinous censure.

Work for WMF?

So very inception-esue.

[The Garbage Scow]

Good old Jdforrester, the lifetime apparatchik-of-many-functions. I miss the Wikitruth page about him.

[The Garbage Scow]

Here's some old-fashioned Moeller-related fun from 2009, wherein some universally despised fundraising banners were disabled and Herr Moeller reversed. More proof of Moeller's dictatorial behavior. I'll just let y'all read for yourselves.

https://en.wikipedia.org/wiki/Wikipedia ... :Eloquence

[Vigilant]

Here's some old-fashioned Moeller-related fun from 2009, wherein some universally despised fundraising banners were disabled and Herr Moeller reversed. More proof of Moeller's dictatorial behavior. I'll just let y'all read for yourselves.

https://en.wikipedia.org/wiki/Wikipedia ... :Eloquence

He seems incapable of learning anything.
That was FIVE YEARS AGO and his approach is just as shitty today.

I noticed that Jehochman is still the servile ass licker that he was back in the day.

[EricBarbour]

Here's some old-fashioned Moeller-related fun from 2009, wherein some universally despised fundraising banners were disabled and Herr Moeller reversed. More proof of Moeller's dictatorial behavior. I'll just let y'all read for yourselves.

https://en.wikipedia.org/wiki/Wikipedia ... :Eloquence

Forgot all about that little incident. Thanks.

And as for Forrester:
https://en.wikipedia.org/wiki/Herbie_Popnecker

[The Garbage Scow]

Here's some old-fashioned Moeller-related fun from 2009, wherein some universally despised fundraising banners were disabled and Herr Moeller reversed. More proof of Moeller's dictatorial behavior. I'll just let y'all read for yourselves.

https://en.wikipedia.org/wiki/Wikipedia ... :Eloquence

Forgot all about that little incident. Thanks.

And as for Forrester:
https://en.wikipedia.org/wiki/Herbie_Popnecker

OMG

[Jim]

And as for Forrester:
https://en.wikipedia.org/wiki/Herbie_Popnecker

Inspired

Back at Operation Deflect All Criticism:

Pawns in Marquis Surcouf are an attempt to retain the character formatting.

...o...k... we added a little chess pawn to be displayed in the article to retain the formatting of the text which was removed. Got it.

and from higher on the page:

The sort key problem needs a new bug report.
The scrollbar problem is a design problem: You can scroll with your fingers (on a laptop trackpad), mouse wheel, or with the arrow keys. If you click on it, it closes without adding a category. Design problem: if "click on it" is repurposed to become the third way to "scroll", then how would you close the search list without adding a category? Whatamidoing (WMF) (talk) 00:36, 29 September 2014 (UTC)

Don't most modern systems show different behaviour when you click on the scroll bar (or its arrows on top and bottom) compared to anywhere else in the box? Thta's rather basic... How would I close the box? With an X in a top corner, or a "close" button, or something else? I know, revolutionary ideas, wil take some time for the WMF devs to get around to them probably. A bit like a special character inserter that let's you add more than one character probably. I'm willing to discuss design issues, but I don't think that "using the scroll bar as a close button" is really a design issue, it's a clear bug. No priority, not assigned... Fram (talk) 08:09, 29 September 2014 (UTC)

All going well, then

[Triptych]

I got a real chuckle out of this, maybe you will too. Though Fram is an admin it doesn't rise to something for the "governance" subforum. A message from Fram to WMF exec. director Lila Tretikov on her Meta talkpage. I don't think I need to set it up more than that, just that I found it humorous maybe you will too. All copy-pasted from: http://meta.wikimedia.org/wiki/User_tal ... of_the_WMF.

Complaint about some employees of the WMF

As an unhappy contributor, I would like to make a complaint about some WMF employees.

* User:Erik Moeller (WMF): he has single-handedly alienated more Wikipedia editors (all languages) than the rest of the WMF combined. I don't think I need to rehash all his problematic decisions, communications, and actions wrt VisualEditor and MediaViewer, which were basically the culmination of years of problems. Please, if you can't or won't get rid of him, then at least make sure that there is no more interaction between Moeller and the projects, and no more communications from Moeller.

* User:Jorm (WMF) is the proverbial example of a software designer who is not interested in what people need, only in what he can make. His work on Flow, and his communications about it, emphasized the gap between his vision and what was actually wanted or needed, and his inability to grasp this and act accordingly. Perhaps he is a good developer, but then make sure that he only produces things for which others have written the specifications, and keep him away from interactions with editors.

* User:Jdforrester (WMF), doing his best to make the exasperation over VisualEditor worse at every possible moment. Together with Erik Moeller and Jasper Deng (not a WMF employee) the main reason I refuse to edit at Mediawiki any more. A very poor communicator when things go wrong, and the manager of a completely botched product launch. See e.g. en:Wikipedia:VisualEditor/Feedback/Archive 2014 3#Testing report from Wbm1058 for a typical passive-agressive example of his lack of clue (also here).

* User:Whatamidoing (WMF) is trying her best, but is just completely incompetent in her work as a community liaison (I go with incompetent, as the alternatives are worse). She is not a community liaison but a crony of the WMF, incapable of seeing problems, incapable of understanding needs, wants, or even basic posts, making up things to defend her (or the WMF's) position, and more often than not a total waste of time. It's always hard to find when things started to go wrong, but where it for me definitely went downhill, never to recover again, was at en:Wikipedia:VisualEditor/Feedback/Archive 2014 1#Feedback request: VisualEditor special character inserter, with unforgettable statements like "I know that some people find the Agile approach irritating, but since you systematically search for incomplete features, you are consenting to it and supporting it. Whatamidoing (WMF) (talk) 23:21, 29 January 2014 (UTC) " Discussions further down that page (e.g. "Empty <ref /> tag added by VE" or "Advanced image settings? More VisualEditor rubbish.") illustrate similar problems with Whatamidoing and JDforrester. These are old, I use them to indicate where things went wrong and for how long these user interaction problems persist already. The current VE feedback page on enwiki shows the ongoing problems with Whatamidoing. See also e.g. en:Wikipedia:VisualEditor/Feedback/Archive 2014 3#Utter waste of money. Another good example is en:Wikipedia:VisualEditor/Feedback/Archive 2014 3#Category additions with VE are FUN!, where Whatamidoingis doing everything but helping. Her attitude is "it is a local enwiki non-VE problem unless you prove to me otherwise", not "hmm, I wonder if this is enwiki only, let's try it out". In many disucssions, she makes categorical statements, which give the impression of her being knowlegeable when they are correct, but only make her look foolish when they are wrong, which happens way too often.

Other editors may have other experiences of course, but for me it really has gone to far for much too long. Fram (talk) 09:31, 25 September 2014 (UTC)

Communication is an item that is on top of my list. I want to say however that we are all trying to work in good faith; nonetheless there is always ways in which we can improve and I -- for one -- am learning ways to do so every day. I don't communicate perfectly, but I work on improving. I can tell you with 100% certainty that am asking the same of everyone at the Foundation as well and now providing resources for doing so. -- LilaTretikov (WMF) (talk) 05:51, 30 September 2014 (UTC)

Thanks. I am not convinced that all of them try to work (for the improvement of Wikipedia) in good faith, at leat one comes across as a power-hungry bully only interested in his own success-rate (product is delivered!), not in the actual result or in what the editing communities want or need; but most of the above are probably good faith workers in the wrong position. Fram (talk) 07:45, 30 September 2014 (UTC)

[Kumioko]

Personally I don't like Fram and I think he should have had his admin privileges removed long ago. With that said, I must grudgingly admit that he is absolutely right on this. I don't think for a second that anything will come out of it with regard to the WMF, but I think he makes excellent points in this case.

[Triptych]

Personally I don't like Fram and I think he should have had his admin privileges removed long ago. With that said, I must grudgingly admit that he is absolutely right on this. I don't think for a second that anything will come out of it with regard to the WMF, but I think he makes excellent points in this case.

Okay, I just thought it was funny he's issuing performance critiques for these employees to Lila.

[Peter Damian]

Okay, I just thought it was funny he's issuing performance critiques for these employees to Lila.

It amused me too, if that's any help.

[thekohser]

Merge this thread to here, please.

Mod note: Merged. --km

[tarantino]

Wow. What a cunt.

To your first point, I do not dismiss them. It's because I don't is the reason I'm careful to say we should be data-informed, not data-driven in our decision-making and why I'm careful to mention that the data is based on reversions and block rate. It's also why the original researcher posted data showing a decrease in productivity post-launch as well as a discussion of the data, instead of "letting numbers speak for themselves." I don't find the features you mention "indispensable," but perhaps this is because I have a different definition of that word based on my technical and engineering background. Personally, as long as wikitext editing exists as an alternative, I'm okay with it—it means the bulk of my editing is still done in source mode. From a technical standpoint, the consequence of omitting a "should have" feature like cut-and-paste in the minimum viable product is a drop in VisualEditor adoption and visual editing productivity. We've seen this. As for getting to the there where we do have cut-and-paste, that, I believe why JamesF has enabled opt-in on enwiki—to bring engineering focus back on working the product and not on these discussions.
To your second point. I believe the statement implies bad faith from a person who has been civil in his dialog and has admitted a number of mistakes where he ended up with compromise or consideration in action with respect to VE: in admitting failure in the way rollout was handled, in his initial stance on a VE opt-out user preference, in the initial position of the tabs, edit links, dialog boxes, and the beta badge, in his re-instatement of opt-in. I don't think anyone can point to a similar record in their actions, and I know I cannot in my own behavior. Given that agreement beyond simple empathy on that particular rant will out me as a hypocrite, you'll pardon me if I disengage and not respond further. :-) - tychay (tchay@wikimedia) (talk) 06:24, 24 September 2013 (UTC)

And you're a director?! Of engineering?!

Not anymore. He's been … off-boarded.

He needs to udpate his linkedin profile, which says,

Director of Features Engineering
Wikimedia Foundation
Nonprofit; 201-500 employees; Internet industry
February 2012 – Present (2 years 9 months) San Francisco Bay Area

I do the stuff on Wikipedia that pisses off the entire community. Actually my team does it, and I just take credit for it.

[Vigilant]

Still a director at meta...

http://meta.wikimedia.org/wiki/User:Tychay

Terry Chay
Director of Features Engineering, Wikimedia Foundation

Apropos

My work
My role is to be administrative overhead at the Foundation.

His homepage is suitably unprofessional and mentions that October 3rd was his last day.
http://terrychay.com/

[EricBarbour]

Not anymore. He's been … off-boarded.

This is unlikely/bizarre, because as I said last year:

"Terry Chay is an insider at Freenode. One of those people who helps to protect the Wikipedia IRC channels. Oh yeah, they'd want him on the payroll."

Perhaps Lila is starting to realize that IRC isn't such a "wonderful asset" to the WMF?

[Kelly Martin]

First victim of the housecleaning?

[Vigilant]

First victim of the housecleaning?

My guess.

Maybe he was scheduled to be moved to the Product side and quit instead...

[Hex]

He's been … off-boarded.

I thought that was going to be you parodying WMF-speak but they actually said it!!

[The Adversary]

That is not the only one, look at logs for JulieT_(WMF): another familiar name who has "off-boardered" is (now former) Product Manager Steven Walling.

[Vigilant]

A familiar face that Allie and I agreed to disagree over.

22:40, 21 January 2014 JulieT (WMF) (talk | contribs) removed the account of "LeslieCarr (talk | contribs)". Duration of the removal: indefinite (User no longer works for the Wikimedia Foundation)

https://wikimediafoundation.org/wiki/User:LeslieCarr

[thekohser]

That is not the only one, look at logs for JulieT_(WMF): another familiar name who has "off-boardered" is (now former) Product Manager Steven Walling.

Wait, does this mean that Walling is no longer an employee, or that the account "Steven (WMF)" is no longer active?

Never mind -- I forgot his jump to Quora.

[Thracia]

Latest rather sparsely-attended IRC office hours on VE (link) explain why those chess pawn symbols unexpectedly kept appearing in VE-edited articles.

Turns out they were part of what WMF Engineering calls a "clever trick", or what normal software developers would call a "dirty great hack".

Good news though! Six months of work has gone into replacing this "clever trick" with much the same "clever trick", but now the pawns will be "invisible unicorns" instead, no doubt in honour of their beloved leader, Purple.


Excerpt (my emphasis):

[18:02:15] <James_F> OK, since September our work has mostly focussed on language support and table editing.
[18:02:56] <James_F> On the first of these, a really major re-write of how we intercept browser behaviour and correct it so that bold/italics/links can be extended when you type next to an existing style.
[18:03:29] <James_F> Different browsers unfortunately do this in different ways; some of them always extend, some of them never extend, some of them do it but only in some conditions.
[18:03:54] <James_F> So a big part of VisualEditor's job is to correct what your browser wants to do so that it works the same whatever browser you happen to be using (and, of course, works at all).
[18:04:23] <James_F> Previously we were using a clever trick we called "pawning", where we put in a pawn character and then removed it again.
[18:05:17] <James_F> Sometimes this broke, especially when a browser changed its mind half way through an event, which could leave you with pawn characters ("â™™") in the document.
[18:05:43] <James_F> So for the past six months or so, we've been working to replace this technology with an alternative that we're calling "unicorning".
[18:06:33] <James_F> This uses an image of a chess knight with an acute accent (so it looks like a unicorn) instead of a character.
[18:07:40] <James_F> This has a huge advantage in how it works, which means that people using Input Method Editors (e.g. for Japanese, Korean, Arabic or Indic language text entry) are a lot closer to being able to use VisualEditor reliably.
[18:08:10] <James_F> Pawning interfered with IMEs so that you couldn't use them at all; with unicorning, we're not done yet but we've made a lot of progress.
[18:08:55] <James_F> Obviously fixing it so all IMEs work flawlessly is a lot of work, but it's one of our major commitments for the next few months and we'll hopefully have more to say on this soon.
[18:09:39] <James_F> The other advantage is that the images don't ever make it into the "model", which means that incidents when editing with VisualEditor suddenly turns into a chess game, as Elitre puts it, should go away.
[18:09:57] <Elitre> James_F: so will editors for those Wikipedias already notice an improvement?
[18:10:01] <James_F> There's still a little bit of pawning left which we hope to replace this week, but in most cases it's gone now.
[18:10:28] <James_F> Elitre: For some IMEs VisualEditor now works where previously it failed, yes, but in general we're still not done with a lot of the major IMEs.
[18:10:45] <James_F> Elitre: E.g. Linux iBus in Malayalam.
[18:10:51] <Elitre> (I'm not sure if no images is a plus. I bet most people wouldn't mind to see unicorns here and there.)
[18:10:59] <James_F> Ha. :-)
[18:11:16] <James_F> Also, the unicorns are invisible in normal operation, just to be boring. :-)
[18:12:05] <James_F> OK, so, that's where we're up to with language support.

[The Garbage Scow]

Latest rather sparsely-attended IRC office hours on VE (link) explain why those chess pawn symbols unexpectedly kept appearing in VE-edited articles.

Turns out they were part of what WMF Engineering calls a "clever trick", or what normal software developers would call a "dirty great hack".

Good news though! Six months of work has gone into replacing this "clever trick" with much the same "clever trick", but now the pawns will be "invisible unicorns" instead, no doubt in honour of their beloved leader, Purple.


snip

Hilarious. So in other words, they can't write something that works as intended, so they're going to go on inserting nonsense characters into the middle of stuff. So if anyone actually still uses VE, the result could be a bunch of articles with invisible unicorns embedded in the wikicode?

Maybe someone can write a bot to go around and remove the invisible unicorns afterwards.

It's Miller time!

[Vigilant]

Jesus fucking Christ!

I hope your next potential employer asks you about this in your interview.

[Kelly Martin]

Latest rather sparsely-attended IRC office hours on VE (link) explain why those chess pawn symbols unexpectedly kept appearing in VE-edited articles.

Turns out they were part of what WMF Engineering calls a "clever trick", or what normal software developers would call a "dirty great hack".

Good news though! Six months of work has gone into replacing this "clever trick" with much the same "clever trick", but now the pawns will be "invisible unicorns" instead, no doubt in honour of their beloved leader, Purple.


Excerpt (my emphasis):

[18:02:15] <James_F> OK, since September our work has mostly focussed on language support and table editing.
[18:02:56] <James_F> On the first of these, a really major re-write of how we intercept browser behaviour and correct it so that bold/italics/links can be extended when you type next to an existing style.
[18:03:29] <James_F> Different browsers unfortunately do this in different ways; some of them always extend, some of them never extend, some of them do it but only in some conditions.
[18:03:54] <James_F> So a big part of VisualEditor's job is to correct what your browser wants to do so that it works the same whatever browser you happen to be using (and, of course, works at all).
[18:04:23] <James_F> Previously we were using a clever trick we called "pawning", where we put in a pawn character and then removed it again.
[18:05:17] <James_F> Sometimes this broke, especially when a browser changed its mind half way through an event, which could leave you with pawn characters ("â™™") in the document.
[18:05:43] <James_F> So for the past six months or so, we've been working to replace this technology with an alternative that we're calling "unicorning".
[18:06:33] <James_F> This uses an image of a chess knight with an acute accent (so it looks like a unicorn) instead of a character.
[18:07:40] <James_F> This has a huge advantage in how it works, which means that people using Input Method Editors (e.g. for Japanese, Korean, Arabic or Indic language text entry) are a lot closer to being able to use VisualEditor reliably.
[18:08:10] <James_F> Pawning interfered with IMEs so that you couldn't use them at all; with unicorning, we're not done yet but we've made a lot of progress.
[18:08:55] <James_F> Obviously fixing it so all IMEs work flawlessly is a lot of work, but it's one of our major commitments for the next few months and we'll hopefully have more to say on this soon.
[18:09:39] <James_F> The other advantage is that the images don't ever make it into the "model", which means that incidents when editing with VisualEditor suddenly turns into a chess game, as Elitre puts it, should go away.
[18:09:57] <Elitre> James_F: so will editors for those Wikipedias already notice an improvement?
[18:10:01] <James_F> There's still a little bit of pawning left which we hope to replace this week, but in most cases it's gone now.
[18:10:28] <James_F> Elitre: For some IMEs VisualEditor now works where previously it failed, yes, but in general we're still not done with a lot of the major IMEs.
[18:10:45] <James_F> Elitre: E.g. Linux iBus in Malayalam.
[18:10:51] <Elitre> (I'm not sure if no images is a plus. I bet most people wouldn't mind to see unicorns here and there.)
[18:10:59] <James_F> Ha. :-)
[18:11:16] <James_F> Also, the unicorns are invisible in normal operation, just to be boring. :-)
[18:12:05] <James_F> OK, so, that's where we're up to with language support.

This is some sort of clever parody and not actually what's going on, right?

[Vigilant]

Latest rather sparsely-attended IRC office hours on VE (link) explain why those chess pawn symbols unexpectedly kept appearing in VE-edited articles.

Turns out they were part of what WMF Engineering calls a "clever trick", or what normal software developers would call a "dirty great hack".

Good news though! Six months of work has gone into replacing this "clever trick" with much the same "clever trick", but now the pawns will be "invisible unicorns" instead, no doubt in honour of their beloved leader, Purple.


Excerpt (my emphasis):

[18:02:15] <James_F> OK, since September our work has mostly focussed on language support and table editing.
[18:02:56] <James_F> On the first of these, a really major re-write of how we intercept browser behaviour and correct it so that bold/italics/links can be extended when you type next to an existing style.
[18:03:29] <James_F> Different browsers unfortunately do this in different ways; some of them always extend, some of them never extend, some of them do it but only in some conditions.
[18:03:54] <James_F> So a big part of VisualEditor's job is to correct what your browser wants to do so that it works the same whatever browser you happen to be using (and, of course, works at all).
[18:04:23] <James_F> Previously we were using a clever trick we called "pawning", where we put in a pawn character and then removed it again.
[18:05:17] <James_F> Sometimes this broke, especially when a browser changed its mind half way through an event, which could leave you with pawn characters ("â™™") in the document.
[18:05:43] <James_F> So for the past six months or so, we've been working to replace this technology with an alternative that we're calling "unicorning".
[18:06:33] <James_F> This uses an image of a chess knight with an acute accent (so it looks like a unicorn) instead of a character.
[18:07:40] <James_F> This has a huge advantage in how it works, which means that people using Input Method Editors (e.g. for Japanese, Korean, Arabic or Indic language text entry) are a lot closer to being able to use VisualEditor reliably.
[18:08:10] <James_F> Pawning interfered with IMEs so that you couldn't use them at all; with unicorning, we're not done yet but we've made a lot of progress.
[18:08:55] <James_F> Obviously fixing it so all IMEs work flawlessly is a lot of work, but it's one of our major commitments for the next few months and we'll hopefully have more to say on this soon.
[18:09:39] <James_F> The other advantage is that the images don't ever make it into the "model", which means that incidents when editing with VisualEditor suddenly turns into a chess game, as Elitre puts it, should go away.
[18:09:57] <Elitre> James_F: so will editors for those Wikipedias already notice an improvement?
[18:10:01] <James_F> There's still a little bit of pawning left which we hope to replace this week, but in most cases it's gone now.
[18:10:28] <James_F> Elitre: For some IMEs VisualEditor now works where previously it failed, yes, but in general we're still not done with a lot of the major IMEs.
[18:10:45] <James_F> Elitre: E.g. Linux iBus in Malayalam.
[18:10:51] <Elitre> (I'm not sure if no images is a plus. I bet most people wouldn't mind to see unicorns here and there.)
[18:10:59] <James_F> Ha. :-)
[18:11:16] <James_F> Also, the unicorns are invisible in normal operation, just to be boring. :-)
[18:12:05] <James_F> OK, so, that's where we're up to with language support.

This is some sort of clever parody and not actually what's going on, right?

You know, Kelly, that exact thought passed through my head for a second.

Pull the other one.

Then I remembered what VE looked like at launch, the feeble protestation that they NEEDED live testing at that point, it's gonna get a bunch better REAL SOON NOW

[Kelly Martin]

You know, Kelly, that exact thought passed through my head for a second.

Pull the other one.

Then I remembered what VE looked like at launch, the feeble protestation that they NEEDED live testing at that point, it's gonna get a bunch better REAL SOON NOW

Not to mention almost exactly the same "strategy" is used at various points in the wikitext parser: the parser generates a "cookie" and global-regexp-replaces certain constructs with the "cookie" to prevent them from being matched by later global-regexp-replacement transformations of the wikitext, then later on replaces the cookie with whatever it really supposed to be.

Repeat after me: preg_replace is a piss-poor way to implement a parser.

[Vigilant]

You know, Kelly, that exact thought passed through my head for a second.

Pull the other one.

Then I remembered what VE looked like at launch, the feeble protestation that they NEEDED live testing at that point, it's gonna get a bunch better REAL SOON NOW

Not to mention almost exactly the same "strategy" is used at various points in the wikitext parser: the parser generates a "cookie" and global-regexp-replaces certain constructs with the "cookie" to prevent them from being matched by later global-regexp-replacement transformations of the wikitext, then later on replaces the cookie with whatever it really supposed to be.

Repeat after me: preg_replace is a piss-poor way to implement a parser.

Maybe they should send someone out for a "Compiler Design" class at a local university?

[Kiefer.Wolfowitz]

You know, Kelly, that exact thought passed through my head for a second.

Pull the other one.

Then I remembered what VE looked like at launch, the feeble protestation that they NEEDED live testing at that point, it's gonna get a bunch better REAL SOON NOW

Not to mention almost exactly the same "strategy" is used at various points in the wikitext parser: the parser generates a "cookie" and global-regexp-replaces certain constructs with the "cookie" to prevent them from being matched by later global-regexp-replacement transformations of the wikitext, then later on replaces the cookie with whatever it really supposed to be.

Repeat after me: preg_replace is a piss-poor way to implement a parser.

Maybe they should send someone out for a "Compiler Design" class at a local university?

Am I the only one whose synapses are misfiring because of the parallel thread on Ullman?

[Kelly Martin]

You know, Kelly, that exact thought passed through my head for a second.

Pull the other one.

Then I remembered what VE looked like at launch, the feeble protestation that they NEEDED live testing at that point, it's gonna get a bunch better REAL SOON NOW

Not to mention almost exactly the same "strategy" is used at various points in the wikitext parser: the parser generates a "cookie" and global-regexp-replaces certain constructs with the "cookie" to prevent them from being matched by later global-regexp-replacement transformations of the wikitext, then later on replaces the cookie with whatever it really supposed to be.

Repeat after me: preg_replace is a piss-poor way to implement a parser.

Maybe they should send someone out for a "Compiler Design" class at a local university?

Am I the only one whose synapses are misfiring because of the parallel thread on Ullman?

It has occurred to me to muse about whether the WMF's unicorns would be a match for a dragon or two.

[Kiefer.Wolfowitz]

[Snowtrooper]

[Vigilant]

Once more into the breach, dear friends, once more
https://en.wikipedia.org/wiki/Wikipedia ... test_of_VE

Gotta get it everywhere.

[Vigilant]

Still major bugs outstanding, still major, obvious bugs being reported, still a 1.3% usage rate on en.wp.

What's the cost per edit made on en.wp?

[Jim]

What's the cost per edit made on en.wp?

I dunno. Too much.
I tried it again yesterday, and it's still no use to me. It splattered <nowiki> tags all over my edit.
I should "re-train" myself in order to use something "easier"? Yeah, no.

[Vigilant]

Keeping up with failed projects...

Using the 5000 most recent changes as a basic yardstick, the much vaunted VisualCowpat is getting roughly 28 of the last 5000 edits day in and day out.

0.56% of the edits made to the largest wiki project are coming from a piece of software being built and maintained, poorly, by how many engineers?
They couldn't be doing something more useful?

[Notvelty]

Keeping up with failed projects...

Using the 5000 most recent changes as a basic yardstick, the much vaunted VisualCowpat is getting roughly 28 of the last 5000 edits day in and day out.

0.56% of the edits made to the largest wiki project are coming from a piece of software being built and maintained, poorly, by how many engineers?
They couldn't be doing something more useful?

That's a good thing, surely? Imagine if it was an actual production piece of software affecting a sizable portion of the edit population, rather than a unused evolutionary dead end.

[Vigilant]

Keeping up with failed projects...

Using the 5000 most recent changes as a basic yardstick, the much vaunted VisualCowpat is getting roughly 28 of the last 5000 edits day in and day out.

0.56% of the edits made to the largest wiki project are coming from a piece of software being built and maintained, poorly, by how many engineers?
They couldn't be doing something more useful?

That's a good thing, surely? Imagine if it was an actual production piece of software affecting a sizable portion of the edit population, rather than a unused evolutionary dead end.

It's a silver lining that it was never KEPT as mandatory.
It's pre-SuperProtect.

[HRIP7]

Keeping up with failed projects...

Using the 5000 most recent changes as a basic yardstick, the much vaunted VisualCowpat is getting roughly 28 of the last 5000 edits day in and day out.

0.56% of the edits made to the largest wiki project are coming from a piece of software being built and maintained, poorly, by how many engineers?
They couldn't be doing something more useful?

[Vigilant]

Good lord, three years ago.

I can't think of a more monumental failure inside the WMF engineering group.

[Anthonyhcole]

Jimmy still thinks fixing the MediaWiki interface will fix Wikimedia's problems.

[Hex]

Jimmy still thinks....

"Omit needless words." -- Strunk & White

[Lukeno94]

Jimmy still pretends to think....

"Omit needless words." -- Strunk & White

FTFY

[DanMurphy]

Some highlights from the blog announcing the relaunch of the "epically important" Visual Editor.

You may remember a similar announcement in December 2011, when we revealed a developer prototype of our “visual editor,” but after a great deal of feedback, we’ve reworked it so that it’s more useful to our community of users.

...To build this iteration of our open-source visual editor, we have been working with some of the team from Wikia, a collaborative publisher that operates the largest network of video game, entertainment and lifestyle wikis in the world. We both believe that this kind of tool should be built not just for the Wikimedia wiki projects, but for everyone using MediaWiki software, and when it’s done we look forward to including the visual editor “out of the box” for anyone setting up a wiki with our software.

And perhaps for a "where are they now?" segment some day.

- The Visual Editor Team: Trevor Parscal, Inez Korczyński, James Forrester, Roan Kattouw, Rob Moen, Subramanya Sastry, Brion Vibber, Gabriel Wicke, Christian Williams.