Trains:Station Cafe (The Works)

From TrainSpottingWorld, for Rail fans everywhere
(Redirected from Works)


Shortcut:
Works

Welcome to the The Works, the area of the Station Cafe dedicated to discussions about the mechanics of the Train Spotting World Wiki, page formatting, etc.


This is one of the Station Cafe pages. We use them to discuss the technical issues, policies, and operations of Train Spotting World. The Station Cafe is divided into sections covering these different aspects, but if you can't decide which to use, try Left Luggage. You can see all station cafe sections at once here. Please sign and date your posts (by typing ~~~~ at the end).

Station Cafe sections
Buffet Car
post | watch

TSW Policy

The Works
post | watch

Technical Issues

Mess Room
post | watch

Ideas and Suggestions

Guard's Van
post | watch

Requests for Help

Left Luggage
post | watch

Everything else


Main Page 'Breaking News' Mechanism

I have highlighted a need for a new mechanism for 'Breaking News', and have outlined one suggestion. This requires consideration by other editors.

Please join the discussion at Trains talk:Breaking News#Defined Mechanism Needed.

EdJogg 12:14, 13 November 2007 (UTC)

Upload PD Images

Drop-down menu on Upload page does not have an entry for 'I am NOT creator, but creator has released image into Public Domain', as for this roadworks sign. Can this be updated?

EdJogg 19:04, 14 November 2007 (UTC)

If it has been released elsewhere under the GFDL licence (Which all WP is after itejms are released into it, except fair Use, of course) then the right licence here is GFDL, isn't it? Source attribution is necessary {{From Wikipedia}} applies (even if it is stored on Commons I think) and then {{GFDL}}
Or am I mistaken? Tim Trent 19:41, 14 November 2007 (UTC)

That little "+" sign

We really need to have that in the set of tabs at the top of this page. It makes it easy to add a new section. I can't find out how to do it. It's on talk pages, of course it is, but not here. If anyone has any idea how, please point us in the right direction. Tim Trent 14:04, 19 November 2007 (UTC)

And solved by pure accident. __NEWSECTIONLINK__ does it for non talk pages. Tim Trent 14:36, 19 November 2007 (UTC)
Well done, Sir! We'll make you a Mediawiki expert yet!!! EdJogg 14:51, 19 November 2007 (UTC)
I would give you the trains barnstar (but probably the newly made Mediawiki barnstar), but I havnt made it yet!!! Bluegoblin7 16:38, 19 November 2007 (UTC)

URL encoding

The template {{PAGENAMEU}} makes use of the mediawiki syntax {{urlencode:FILENAME}}. However, this function replaces space characters with '+' symbols. When such a link is followed, the destination page continues to use '+' symbols in the address, thereby preventing the page from being found. If a URL containing spaces is cut-and-pasted to the address bar, the spaces are auto-converted to underscores, and the URL is 'understood'. This is why the {{From Commons}} template displays the URL without providing a direct link (there is no Interwiki link from TSW to Commons). Solutions to providing a direct link would be welcome!

EdJogg 14:14, 19 November 2007 (UTC)

Have a look at Help:Magic words and see if that solves your problem. Tim Trent 14:33, 19 November 2007 (UTC)
I've got you half way plus! I've commented out your first part and entered modified stuff from {{From Wikipedia}}. The interwiki linking works, but I can't work out the syntax of the conditional elements for "if it's the same pagename, then use it, else use what I write here", so I've not handled that.
When you handle it in {{From Commons}} please could you, after testing it works, reverse engineer that into {{From Wikipedia}}?
Tim Trent 16:04, 19 November 2007 (UTC)
I keep staring at that pesky "if" statement and wondering how to add it to either of these templates to idiot-proof them. This multiple {{{{{{}}}}}} syntax is alarming. I have, so far, chickened out at every available opportunity. I believe in consistency! Tim Trent Talk to me 16:17, 22 November 2007 (UTC)
Sorry, I keep getting side-tracked by everything else! This shouldn't be too difficult to do (well, for a software engineer :o) ), just need to put my mind to it. If you want a really 'serious' template to look at, try the source code for the Trains WikiProject banner (written by template-meister Slambo).
EdJogg 17:06, 22 November 2007 (UTC)
I should have been doing something much more important, but it's done.
{{From Commons}} is 'complete', apart from having the documentation section moved to a sub-page. You will see that the named parameter has been dispensed with, as I have worked out an alternative solution for ignoring the 'en' parameter on {{Used}}.
{{From Wikipedia}} has been upgraded to match. Can now place a parameterless template on the talk page and it will link to the correct WP article. The only drawback to this approach is that it assumes article namespace, but I didn't think this was a problem since we have been liberally copying templates without noting the source! (If we need to, a new Template:Template from Wikipedia will be appropriate.)
I have documented {{From Commons}}, but I'll leave you to apply the {{From Wikipedia}} documentation, and update any usage notes elsewhere...
EdJogg 23:11, 22 November 2007 (UTC)

Archiving

Wikipedia has an excellent set of archiving bots which archive talk pages and other pages (Such as their Village pump). We will need such things here. Anyone fancy taking the job on of identifying the right one and porting it across here? Tim Trent 14:38, 19 November 2007 (UTC)

We could probably do with a help page on archiving. I think we already have all the templates necessary for the talk pages, but it would be nice to be able to guide people how to use them effectively.
EdJogg 14:53, 19 November 2007 (UTC)
Probably right, but almost no-one implements a manual archiving scheme. I think we need a bot, too, then we can produce the right help. My own talk page is archived and examples of the use of templates can be seen there. Tim Trent 15:08, 19 November 2007 (UTC)
Will have a look at some of the archiving bots on WP, and see if they could be implemented here. Will also speak to owners. Bluegoblin7 16:49, 19 November 2007 (UTC)
I've used Shadowbot3 since Werdnabot went offline. It seems robust and reliable, which is all we need. I appreciate your taking this one on. Tim Trent 17:08, 19 November 2007 (UTC)

svg files

These specialist, xml defined images are sometimes displaying and other times not. the issue seems to be that some svg editing suites put tags in that Mediawiki extensions fail to recognise. This is a consistent error.

We've identified it and are working on a resolution. So, continue to upload svg files, even if they are invisible for now! It's "one of those things" that will suddenly come right when we finish the resolution.

Tim Trent 14:42, 19 November 2007 (UTC)

That one is solved. We found the problem was with files created by Inkscape so installed it as a renderer. It renders svg as png (all because M$ was too slow obeying the emerging svg standard) so we then had to delete all the pre-rendered .svg.png files. Now it all works. Tim Trent 22:22, 19 November 2007 (UTC)

Trains:Station Cafe (all)

I am not sure how to handle the TOC (to get it to indent one more step) and the oft repeated initial template. So I'd appreciate someone else taking a further look. Tim Trent 14:47, 20 November 2007 (UTC)

The indented numbering scheme was just obvious with a little lateral thought! Still not sure what to do with the repetitive template, though. Tim Trent 15:28, 20 November 2007 (UTC)
Try sticking <noinclude> around it... EdJogg 16:00, 20 November 2007 (UTC)
Hiding in plain sight, that was! Thanks. Tim Trent 16:05, 20 November 2007 (UTC)

Protection categories

This isn't very urgent, but having just protected the Station Cafe portal, I have managed to add it to two categories: Category:Protected (added manually) and Category:Semi-protected added by the template...

Now the problem is not that, but rather that the protection categories and templates are in a bit of a mess. (It would help if protecting a page automatically added it to a category, or automatically showed a little padlock, or both).

{{sprotected2}} adds the page to 'semi-protected', but it is the only one of the protection templates that I trust to behave.

What I think I am trying to get at is that there should be different templates for protected (sysop-only) pages and semi-protected (registered users only) pages. Something to add to the site-wide to-do list, maybe?

EdJogg 00:20, 21 November 2007 (UTC)

{{Protected2}} is ready and meets some of that need. I guess we just need to se what we have and bring (and modify) others in from WP.
On that subject I think imagemap now works (has been highly relevant to imported templates). I've discovered the really banal reason why it's needed! Mediawiki wiki markup means that an image cannot be a wikilink except to itself. So the imagemap has to be declared if you want an image to wikilink elsewhere. Someone designed this stuff!
Tim Trent talk to me 14:52, 21 November 2007 (UTC)

FAQ

I have got deep into the mud at Adding an event to the calendar, partly because my laptop is on a go slow and I have to go offline for a few hours. Either some kind soul could have a look or I'll solve it when I get back (I hope). I was using the "All Station Cafe" page as my model and hit a brick wall. What I was attempting is a template for all FAQ pages that is included in the FAQ individual page, but that is not included in the overall FAQ list. One thing I see straight away is that all individual FAQ pages need a title in order to be editable form the FAQ list. Tim Trent talk to me 09:00, 21 November 2007 (UTC)

Well, for a start you were missing a slash! Now inserted. Any help? EdJogg 13:55, 21 November 2007 (UTC)
Have you noticed that, when you give yourself plenty of time to do something but have a deadline you must meet to finish, that there is always something that makes you miss that deadline. That slash solved it. Thank you.
I think you are concerned about the "bigness" of headings created with a single "=". We may be able to use CSS to solve that one.
Tim Trent talk to me 14:46, 21 November 2007 (UTC)

Portal Temps...

Can anyone have a look at the Trains:Adopt-a-User area? The page is screwed up. It all comes from Portal:Box-footer, which I can't get from WP cos it gives me a strange link! HELP!!! Bluegoblin7 15:58, 23 November 2007 (UTC)

I'm guessing that it is mismatched <div></div> tags like in templates. Hard to spot I fear. Tim Trent Talk to me 16:24, 23 November 2007 (UTC)
In Portal:Box-footer there was one too many </div>, but removing it (I have) makes no difference. The problem must be higher up. Tim Trent Talk to me 16:28, 23 November 2007 (UTC)

Work in progress

I have done as follows:

<nowiki>{{Trains:Adopt-a-User/box-footer|}}</nowiki>

throughout

Note: at first sight this cures the problem, but it does not. The left margin is wrong - look at the right hand edge of the SEARCH box.

I've run out of time to play for the moment. My suggested technique is to use <nowiki></nowiki> to surround things to eliminate them. I suspect the box-header, and possibly overall table widths if not in relative measure.

I also don't understand the parameter passed to Portal:Box footer.

Trains:Adopt-a-User/box-footer shows the search box damage, too, but Portal:Box-footer does not. Hmm.

EdJogg? You seem to have a methodical way of solving this type of horror?

Tim Trent Talk to me 17:01, 23 November 2007 (UTC)

BTW Firefox and IE have radically different lpage renderings! Safari shows a subtle narrowing of the left hand menu margin, so it's hard to see the problems. Ye gods and little fishes. What one can "get away with" on WP one can not here. Tim Trent Talk to me 17:25, 23 November 2007 (UTC)
I am more than a little at a loss. I copied a great swathe of CSS from WP to MediaWiki:Common.css that was possibly the right stuff, and, of course, nothing happened. I can't see on WP the class "noprint" which might be the problem, so sorry, I am still very much in the dark here. Tim Trent Talk to me 21:10, 23 November 2007 (UTC)

Partial solution

You imported, or created Portal:Box-footer and Template:Portal:Box-footer. There is an order of transclusion precedence. It goes to the Template namespace first then other namespaces. There were mismatched div tags in each. I cured the former, but not the latter originally.

Now both have matched div tags. But which one do you WANT? Please delete the other one.

But we are not completely there yet. User:Tim Trent/Sandbox has it working for just that segment. Trains:Adopt-a-User is still not right. It;s better, but wrong still. But I think I can safely leave more detective work to you on the basis of teaching you to fish instead of giving you a fish?

Seriously it is worth getting to grips with this now I've aimed you in the right direction.

You can see what "Templates are in use" when opening a page in edit mode and scrolling down past the edit box.

Let us know how you're doing!

Tim Trent Talk to me 23:02, 23 November 2007 (UTC)

BTW a good tip is to debug each transclusion in a sandbox one at a time. If the footer was wrong I wonder about the header. Idle wondering and I have not checked. Tim Trent Talk to me 23:08, 23 November 2007 (UTC)

Solved?

Trains:Adopt-a-User had an unclosed comment towards the bottom, which was hiding a 'div', thus giving the effect of too many /divs. This restored the LH toolbar. Was that the problem you were having?

(BTW - Took less than 5mins to spot and fix, but then this is the sort of thing I do for a living! -- looking for software bugs, I mean, not editing TSW pages!!)

BTW2 - {{Portal:Box-header}} looks a bit scary! Please don't introduce any bugs into that!!!

EdJogg 03:09, 24 November 2007 (UTC)

They all look scary! Looked pretty hard to spot to me. Wheeeeeeeeeeeee. Tim Trent Talk to me 07:52, 24 November 2007 (UTC)
No. Doesn't seem to have worked. The layout is fixed, but the writing should be inside the coloured background. See the WP Page to see what i mean!
This may be something to do with the CSS class used by WP. I suspect the templates work just fine but the class is absent. I'll do some leisurely poking about over the next few days. Work on the text etc, importing other templates etc should just go on, I think. Tim Trent Talk to me 16:19, 24 November 2007 (UTC)
First pass says I am wrong. I can't spot where the box itself is given a background colour. UNoless, of course it is in the column templates? Not looked yet. Tim Trent Talk to me 17:10, 24 November 2007 (UTC)
Sorry, can't see anything either. EdJogg 00:59, 25 November 2007 (UTC)
Somewhat bizarre, really. Do we know anyone who might take a look? Slambo hasn;t been active here for a while, he might be able to spot it easily. Tim Trent Talk to me 07:49, 25 November 2007 (UTC)
Well, you could always ask him to drop by. He's very busy at WP usually, but he might be interested enough to help. (After all, the worst he's likely to say is 'No'.) EdJogg 21:50, 26 November 2007 (UTC)
Yeah, well, I meant at WP, actually...y'know, like where he might see it?  :o)
EdJogg 22:40, 26 November 2007 (UTC)
I did. As did BG7, but I left the main message on his local talk page :) After all TSW stuff is "out of place" on WP Tim Trent Talk to me 22:48, 26 November 2007 (UTC)

Imagemap and svg files

I am starting to hate the alleged need for imagemap, a need we inherited from WP.

We "need" imagemap because it's impossible to use an image in mediawiki (as fas as I know) as a link to another location unless it is bounded by an image map. Images link to themselves, not to other places. So WP deploys imagemap as a means not to create a real image map (bounded areas of an image which each perform a different function when clicked), though it could, but as a means of getting an image to be clickable at all as a link.

Thus syntax like:

<imagemap>
Image:Vista-Ym.png|110px
rect 0 0 128 125 [[Trains:Welcoming committee|The Welcoming Committee]]
desc none 
</imagemap></center>
produces
The Welcoming CommitteeVista-Ym.png

which is a clickable image.

Replace Image:Vista-Ym.png with Image:Peace dove.svg and a different issue arises! SVG files are true images, but not images on which (currently) imagemap may be deployed:

<imagemap>
Image:Peace dove.svg|110px
rect 0 0 128 125 [[Trains:Kindness Campaign|Kindness Campaign]]
desc none 
</imagemap></center>
produces
Kindness CampaignPeace dove.svg

Or an image that is unclickable.

I love the foibles of Mediawiki!

Tim Trent Talk to me 08:35, 25 November 2007 (UTC)

Hi Tim,
It is actually possible to use a template to do this with ALL image types - i'll port it over when i remember where abouts it is!
THanks,
BG7 10:21, 2 April 2008 (UTC)

Updating Image Files

The ex-Commons template {{Used}} has been adapted to identify images sourced from Commons. I have since changed it to add the images to Category:Images sourced from Wikimedia Commons. However, although the images all seem to include the new cat, the cat doesn't include the images! I edited one image slightly, and this now appears, but there are 63 images using this template and I don't want to edit them all.

How is a change to Category:Articles sourced originally from Wikipedia applied to all the files when it is updated in the same way? It seems to work by 'magic'!

(I've just tweaked {{From Wikipedia}} to list articles by pagename. I am currently tweaking the cat, and every time I edit it, more and more articles have been sorted... So why didn't that work for the image files for {{Used}}?)

EdJogg 09:34, 19 November 2007 (UTC)

It seems to have caught up now! EdJogg 10:46, 19 November 2007 (UTC)
What happens is that it has to go and physically update every database record with the new category. It places the task in some sort of job queue and processes these reasonably fast or unreasonably slowly, depending upon your viewpoint :) Tim Trent 11:47, 19 November 2007 (UTC)
I know there are some issues (at WP) about pages not being updated with revised template contents, until the pages in question are physically updated in some way. (Hence the question.) Once the process started, I could see that it was happily trundling along. There are so many files in the 'from WP' cat it was rather intriguing to see a different set of files every time I hit 'Preview'. I was allowing for the differences in speed between the WP server city and whatever we have here at TSW. I guess the job queue on the 'from WP' task was so long that nothing would be seen on the 'from Commons' task until the other had finished.
EdJogg 13:33, 19 November 2007 (UTC)
It's the scale of the brute. We have a pretty minute server compared with the WP server farms, but we have far less activity. Our speed appears to be much faster than WP, which is a hoot!
I think the job queue is wise and interleaves things. Some parts of Mediawiki are great! Itl;s just hard to find them!!
Tim Trent 13:53, 19 November 2007 (UTC)
No criticism was implied. EdJogg 14:03, 19 November 2007 (UTC)
 :) I just loathe Mediawiki, though. Tim Trent 14:09, 19 November 2007 (UTC)

What happens when you update a template

Actually it happens when you update any transcluded page, but templates are the most common.

Edit a template and save it and the system puts all the necessary page updates (Pages which transclude the template) into a job queue. Have a look at Special:Statistics to see the length of the queue at any time.

Any sane designer would then run a cron job that looked at load factors and emptied the queue fast in low load times and slowed down in high load times. Not mediawiki's designers, oh no.

Mediaiwki's job queue empties based upon surfing activity. "By default, each time a request runs, one job is taken from the job queue and executed."

This means that, the more pages viewed by a visitor, the more jobs are called from the queue. One view one job! So Mediawiki adds load to high load period!

Why am I telling you this?

Simply so you know what happens in the background. The queue can take as long as it likes to empty, and large changes will ripple through the database reasonably fast if we have a lot of visitors. No visitors = no queue emptying, but that actually does not matter at all, because they don't need to see anything if they aren't there.

The system is insane, though!

Tim Trent Talk to me 23:21, 28 November 2007 (UTC)

This may amuse you. Tim Trent Talk to me 09:03, 2 December 2007 (UTC)

using the double {

Hey Tim,

Is it possible to transclude a page using the above without havin to have template in front? I can do it at wp, but it won't work here... it's on the main page redesign...

Bluegoblin7 22:12, 30 November 2007 (UTC)

I need an example of where it fails to work. Look at ghe source (currently) of User:Tim Trent/Sandbox for example of it working. Tim Trent Talk to me 09:08, 1 December 2007 (UTC)
It's all on the Main Page Sandbox... Bluegoblin7 17:16, 1 December 2007 (UTC)
FIXED IT!!! (Put a : in front) Bluegoblin7 17:19, 1 December 2007 (UTC)

Template:Random portal component

Hi Tim,

On the main page sandbox, this temp isnt working. Any ideas?

Bluegoblin7 17:52, 1 December 2007 (UTC)

This should be at The Works!
This has got me baffled too. 'Template:Random...' content is same as Wp, so it may be down to the header/footer templates.
Just got a major clue though, while editing the sandbox I just noticed that 'Mod' and 'Rand' are being recognised as templates (which don't exist, of course) rather than the parser functions that they are. Not sure what the answer is, I'm afraid, but you could try testing each of the template components in isolation, gradually building up to the full expression.
Just a thought, Tim: is it possible that TSW S/W doesn't support 'maths' functions like mod and rand? If it doesn't, then "all bets are off"!!
EdJogg 01:20, 2 December 2007 (UTC)
The answer to all questions about the underlying disaster area that we know as Mediawiki is that I have no idea! However WP has wp:Template:Mod and wp:Template:Rand which I have not got the chance right now to port across. Perhaps after one of you guys ports them the thing will work. Of course it all depends on the order in which meediawiki calls things! Tim Trent Talk to me 09:01, 2 December 2007 (UTC)
Ive bought the temps across, but theres no difference. Maybe i'll just make do with a non-randomised one, with several DYKs on it...Bluegoblin7 09:53, 2 December 2007 (UTC)
Don't give up that easily :)
have you done precisely what the instructions say at Template:Random portal component? Sometimes they are written so that they need to be read four or more times!
Tim Trent Talk to me 10:18, 2 December 2007 (UTC)
It's just decided to work! Bluegoblin7 10:23, 2 December 2007 (UTC)
One problem creates another. Now it wants all the subpages to be templates. And the solution above doesnt work. Grr. MediaWiki I hate YOU!!! (And how comes it works on WP ok?)
Bluegoblin7 10:37, 2 December 2007 (UTC)
You may be in error about it needing the sub pages to be template. Any page can be transcluded, but MW assumes a transcluded page is a template because iut goes to the template namespace first. Don't let this beat you. You're good enough to solve it. Tim Trent Talk to me 10:46, 2 December 2007 (UTC)

Ive created all the subpages but no avail. Perhaps its on the job q Bluegoblin7 10:48, 2 December 2007 (UTC)

Add it to the current main page and see if it works/fails there. Tim Trent Talk to me 11:34, 2 December 2007 (UTC)
If that fails EdJogg often has good insights into arcane template syntax. Tim Trent Talk to me 11:35, 2 December 2007 (UTC)
Doesnt transclude on the main page either: but, as it does on the sandbox, you can edit the "real" page through the edit link. Bluegoblin7 18:03, 2 December 2007 (UTC)
I've just added some 'direct' calls to the Main Page, using transclusion syntax, and you'll see that none of them are working. Hence the 'random' template may well be OK, but it will not behave as intended until the pages may be transcluded directly. Back to Tim...?
BTW, it is NOT working in the Sandbox either.
EdJogg 10:07, 3 December 2007 (UTC)
By working in sandbox, i mean the edit link, not the transclusion!
On the main page, I have added colons (see above discussion), and the links now transclude. This, however, doesn't work for the random thingy. What i think we should do, is test it further on Random portal test, and just you use a single DYK box until we can get it working.
I am almost certain it it mediawiki orientated.
Maybe Slambo can shed some light?...
Bluegoblin7 13:22, 3 December 2007 (UTC)
Pure transclusion works, whether it be of articles or templates. The randomisation may be missing something, but I have no idea what. You find it, we'll implement it. I know this is frustrating. Why not start by asking the WP template creator to be very kind and take a look? Tim Trent Talk to me 14:01, 3 December 2007 (UTC)

(reindent) I've just moved the colon to the template, and it appears to be working. I have modified the test call on the Main Page and at Main Page/Sandbox to suit. Perhaps you can test it more thoroughly, etc.

Those subtle TSW vs WP differences again! Think this ought to go in the Templates FAQ somewhere...

EdJogg 14:54, 3 December 2007 (UTC)

Cheers EdJogg...but now we have another problem...Now we can't edit the transcluded pages!!! Bluegoblin7 16:50, 3 December 2007 (UTC)
Is that actually a problem? It is normal practice to include protected content on the Main Page as it is so visible, and hence risky to vandalism.
Incidentally, I don't understand what I'm seeing on the Main Page. The random element appears to say "This is file /Did You Know/1 put in as a test for Bluegobin7 ", yet if you follow the edit link, the text shown is 'test edit'. Is that the problem you meant?
EdJogg 18:15, 3 December 2007 (UTC)
WHat it is is that the page wont transclude correctly, but the edit link is perfect! THe translusion is just /Did You Know, which becomes an external page at http://Did You Know, or similar, and as soon as i press save after editing, my browser closes. I tried to redirect it, but the same happened... grrr... I think i will drop the original maker a message!
Bluegoblin7 19:45, 3 December 2007 (UTC)
Everything is working now... must have been in a mediawiki job q!
But..... the job queue was zero length and has been for ages. Tim Trent Talk to me 20:40, 3 December 2007 (UTC)
(edit conflict)
or not then...
Half fixed. It's still got the URL problem, but at least its showing up now... Bluegoblin7 20:41, 3 December 2007 (UTC)

The Main Page and its boxes

I was going to suggest this on my own talk page, but that woudl not get it discussed properly. Currently, today, there si a problem with boxes and borders on the main page. The border does not surround the text.

Here is my putative solution:

What I think the problem is

Arrghhh. I used to be a programmer in Algol, but this stuff is somehow closing my brain down.

  • We have a sensitivity to mismatched <div></div> tags. They break the finished, saved page layout badly. Try it in a sandbox and see!
  • To solve this on {{/Box-header}}, which oddly appears to be a template that includes {{Portal:Box-header}} which also appears to be a template(!!!), div tags have been matched, and the global page l;ayout issue goes away. The box footer is presumably equivalent.
  • The problem has not been solved, but the symptoms have been masked. Each individual page displays well, but the totality is screwed up.

Now I was thinking about this in bed last night, sad person that I am. And I believe we have twin problems. One ought to be simple to solve, the other requires brain power

Simple issue

I may be mistaken, but I think we need to ensure that {{/Box-header}} is not a template, but a page. Same with {{Portal:Box-header}} (and the footers). This may well mean deletion of what we have, and re-import the correct pages from WP into the correct namespace.

Regular pages transclude, and so do templates, but I think there is a priority order that the Template: namespace is checked before the article namespace.

Let's solve this part first. But please could an expert or a very patient editor who is not distracted do it!

Complex issue

You'll need to bear with me! I'll explain my thinking as best as I can.

WP appears more tolerant of mismatched DIV tags than we are. I think this is because our default layout includes placement of the adverts that pay for this place.

When we ported the relevant items for box headers and footers from WP, their pages had wrecked layout and so did any pages where they were applied. We solved the symptom, but not the problem by matching div opens and closes manually. This was an error. I found this when looking at closing an AfD manually at User talk:S.C.Ruffeyfan where the closure heading had "box border stuff in a div" and the closing div had to be entered manually at he foot of the div to surround the text that needed a border.

Now my thoughts for box-header and box-footer. I think it revolves around <noinclude> tags

Consider this:

putting a border around text
<div (and loads of fancy stuff for a border) >
text text text text text text text text text text text
text text text text text text text text text text text 
text text text text text text text text text text text 
text text text text text text text text text 
</div>

That puts a border around text.

Make the opening and closing stuff into transcluded items

Header:

<div (and loads of fancy stuff for a border) >

Footer:

</div>

Now each of these, when applied thus:

{{Header}}
text text text text text text text text text text text
text text text text text text text text text text text 
text text text text text text text text text text text 
text text text text text text text text text 
{{Footer}}

Create a correct border around text. But their own pages have screwed formats.

What I think we need to do is this:

GoodHeader:

<div (and loads of fancy stuff for a border) >
<noinclude>
</div>
</noinclude>

GoodFooter:

<noinclude>
<div>
</noinclude>
</div>

These have closing and opening div tags respectively that do NOT get transcluded. So their pages will be (should be) OK, and the pages using them, provided they are used in pairs, should be ok.

That's the theory. Would someone sense check this, please?

Tim Trent Talk to me 20:39, 13 December 2007 (UTC)

Discussion

Yes, I agree with the issue regarding the template instead of a page for the box-header. I did notice this but thought about whether this could be an affecting factor. If no one here can fix this, then I know some experienced coders on the english wikipedia who might be able to help. Tbo 157 21:35, 13 December 2007 (UTC)

I think the twin issues are not linked, but I would prefer to solve them "in order" so that we get back to "the way it ought to work". What we have now feels like a solution to transclusion symptoms, not to transclusion problems.
While we are solving this, whoever solves it, and WP coding experts are more than welcome, we need to revert temporarily to the prior Main Page, i think (otherwise we wil get into one heck of a mess! Tim Trent Talk to me 21:50, 13 December 2007 (UTC)
Ive restored the original main page as you said. Tbo 157 22:06, 13 December 2007 (UTC)
Fancy having a crack at the rest of the stuff? I'm far too tired tonight! Tim Trent Talk to me 22:14, 13 December 2007 (UTC)
Forget the 'simple' issue for now. Just go for the 'complex' issue solution as you've outlined it. I suspect you will find that this works fine. When I first saw these two templates, that was exactly what I was expecting to see inside -- split divs. Logically this is needed, or each template nicely completes its own box. (Don't understand how WP gets away with split divs though!) And, as you suggest, making use of noincludes will sort the display issues on the templates' own pages.
As for the 'simple' issue... If the above fixes the 'complex' issue, do check that a 'simple issue' actually exists before changing anything else!
EdJogg 00:10, 14 December 2007 (UTC)
I have no ability to concentrate on this for the next couple of days. I have work pressures, and, far worse than that, family coming over the weekend! I'd be grateful for one of you guys to pick it up and run with it, please. Tim Trent Talk to me 07:19, 14 December 2007 (UTC)
The boxes are solved. See my talk page. It seems my helicopter view was right, but it took sharp eyesight to spot it at ground level. Way to go! Do drop by my talk page for "next actions" coz I'm on big painkillers and can't focus for as long as I want.Tim Trent Talk to me 17:27, 15 December 2007 (UTC)

Discussion 2 - 'Simple Issue'

Just re-read the description of the problem. Please bear in mind that the reason it is weird here is that it is weird at WP, and we copied from one to t'other! I thought about changing the naming at the outset, but the whole setup seemed to need the non-obvious naming, or at least the hierarchical relationship between the various pages.

Also, I would have expected that pages and templates would be transcluded equally. If you have same-name pages in multiple namespaces then you are asking for trouble. I don't think we have that problem here. If anything, at WP, pages will be moved TO the template namespace rather than remaining as mainspace sub-pages.

(BTW - I think that this is actually the more complex issue!!)

EdJogg 00:25, 14 December 2007 (UTC)

From Tim Trent's talk page

I have migrated all the recent general topics from Tim Trent's talk page to here. Thanks. Tbo 157 23:06, 18 December 2007 (UTC)

Interwiki map

Should this project use the interwiki map which can be seen here - [1]. This would allow us to be able to interwiki to many other projects which are part of the map using the same prefix. Thanks. Tbo 157 20:07, 18 December 2007 (UTC)

I don't see why not - can they link to us though?... BG7 20:10, 18 December 2007 (UTC)
PS Are there any other useful MW gadgets we could deploy?
Yes they can link to us. It needs to be done by a meta admin though. What do you mean by gadgets? Archive bots might be useful here. Tbo 157 20:13, 18 December 2007 (UTC)
Well yes: any bot we can use, archive, delivery, sign(!) perhaps. By gadgets i meant thinkgs like the events calendar. Are there any more that might be useful?
BG7 20:16, 18 December 2007 (UTC)
Yes, we could just use the code they use at WP for the bots. We might want to flag them too so well ask Tim to do that. Tbo 157 20:20, 18 December 2007 (UTC)
I spoke to Tim a while ago about bots and he said go for it! We've just got to be careful about what they use.
Another one to (maybe) grab - the Assessment one. I'm probably going to work on it tommorrow as we have several Workforces now...BG7 20:23, 18 December 2007 (UTC)
WEre going to need to ask the operator of the bots for permission fi we can sue the bots and wed probably need to ask them to set it up. And then someone here can run them. Tbo 157 20:25, 18 December 2007 (UTC)
Yes that's probably best - can you do that? Thanks BG7 20:27, 18 December 2007 (UTC)
Ye, sure. Tbo 157 21:15, 18 December 2007 (UTC)

(reset indent)
We do need someone who is experienced in the deployment of bots. I am very nervous of a bot's ability to run wild, too. Testing them on live data is all we have though, whcih is scary.

A bot at WP is GFDL licenced isn't it? We woudl do well to ask the author for help, but GFDL is GFDL. And politeness is also important.

Note: we are unable to use IRC as a communication mechanism between bots.

Should this conversation not be migrated to the cafe?

Tim Trent Talk to me 21:24, 18 December 2007 (UTC)

I have a feeling that not many users would be willing to come here to set up bots unfortunately. Tbo 157 22:26, 18 December 2007 (UTC)


MediaWiki:Common.css

The validator points out alot of errors in the file MediaWiki:Common.css. I know some parts are not recognised by the validator but can I check over it using the wp one as a template? I'll rollback if I screw up. Thanks. Tbo 157 14:25, 18 December 2007 (UTC)

Everything with care :) Do also check out the special page for versions to see what it is we run. We are almost never going to upgrade the mediawiki software unless there is a showstopping problem. It's evil and fragile stuff. Tim Trent Talk to me 15:37, 18 December 2007 (UTC)
Well Wikipedia employ developers and software programmers so they can afford to upgrade etc but im presuming that we don't have that. Tbo 157 16:10, 18 December 2007 (UTC)
Success. Ive managed to hide the main page title using css. Tbo 157
Ive also managed to insert the text "From TrainSpottingWorld, for Rail and Steam fans everywhere" under the heading. I can change it if you want. Is it too long? Tbo 157 17:00, 18 December 2007 (UTC)
Looks good to me :)
What we have is a couple of friends who put this together, half as an academic exercise, and half for fun. Bernd is a consummate programmer and developer, but, quite reasonably, pushes back hard on any change he sees as potential disaster! I have huge enthusiasms that he, wisely, pours cold water on :) And we have found, together, that mediawiki is the most amazing mess of apparently uncontrolled and barely documented collaborative development. Open source it may be, but a good open source project it is not! As an example just look at the Job Queue! We used that as a kluge on servers we had years ago when we had no access to root levels.
We're very happy that any pragmatic changes are made. We'll resist any we think are harmful, but it is hard to harm code one has no access to :)
Our software admin system rolls any extensions out to all three wikis automatically. That does not include local CSS files etc, which must stay local. Tim Trent Talk to me 17:23, 18 December 2007 (UTC)
Yes, Mediawiki is a mess and Im amaazed at how far you've got. It is after all free software. Im presuming that Bernd is a busy person looking at his activity level. So I guess well just have to solve coding problems as a community. Problems Ive come across includes bugs which alter the layout of a page, transcluding box-headers in portal style layouts and dates being changed when a page is deleted and restored. Talking fo bugs, Ive moved the page I created to Sworld:SpottingWorld:Bugs and problems. Tbo 157 17:43, 18 December 2007 (UTC)
We got this far with a lot of cursing :). We chose mediawiki because of the familiar user interface, but frankly hate it as a "product". We also took the decision at design stage that we woudl never touch MySQL, so we use PostgreSQL. BTW, you'll very rarely find Bernd editing here, he prefers the behind the scenes challenges.
We can solve all genuine bugs. One of the challenges can be that a bug to us is sometimes a "tolerance" to WP. The transclusion for things like {{/doc}} there requiring {{page's name/doc}} here may be some sort of patch at WP. Very wary of applying those, but we can if they are showstoppers. If we can work to identify them we can work to solve them, if soluble. The main thing is to document them well enough so that Bernd can see what the issue is when you ask me to draw them to his attention.
Should this conversation migrate to the cafe? If so please cut and paste :)
Tim Trent Talk to me 21:19, 18 December 2007 (UTC)

What else do we allow on this wiki?

Hi Tim,

Just wondering what else is allowed on this wiki either in abscence of the others or here until this one gets split. My ideas were buses, cars etc.

Please let me know,

Bluegoblin7 18:53, 17 December 2007 (UTC)

I guess we allow on this wiki what the community wants. No clear boundaires. Remember TSW is what you want it to be as it says on the main page. Tbo 157 19:17, 17 December 2007 (UTC)
Ok. I suppose the easiest way is to just create Workforces for them, and then if this wiki is spilt, the articles can be moved. They seem more appropriate here than at PSW anyhow. Bluegoblin7 19:19, 17 December 2007 (UTC)
Ye, go for it. Tbo 157 19:52, 17 December 2007 (UTC)
Waiting in the wings we have Trucks, but we can't add it to this server or we'll kill it. But you are right: this is truly (within logical limits) what you want it to be. Tim Trent Talk to me 15:39, 18 December 2007 (UTC)
Ok then. I'll hold back on road transport as that can all go into trucks for the time being. Indeed: WP:Bueses is a child of WP:Trucks.
Any idea when we might get a new server?
Cheers,
Bluegoblin7 15:41, 18 December 2007 (UTC)
The new server depends on the success we have in attracting visitors who click our adverts. To justify a new server the SW community has to generate a repeatable $125 per month in advertising revenue, but that just moves us from a virtual server to a physical server of the same horsepower. It also allows a small war chest to be built up against revenue shortfalls during slow periods. Regrettably we are not yet anywhere near that. Now I truly do not mean "click the adverts" because what we need is a predictable revenue stream. Also any "click campaign" is against Google's terms of service and will get us barred. It stops click fraud, which is a good thing. Tim Trent Talk to me 15:45, 18 December 2007 (UTC)
Ok Tim. I will make sure i do make the odd click if something is relevant - i normally ignore them! I wont go OTT though :-D. Can you let us know when we're nearly there?
Perhaps in the mean time we shoule build up Truck/Bus/Car etc articles in SW? Then they can be moved across. Let me know!
Bluegoblin7 15:50, 18 December 2007 (UTC)
Well, thats the problem with most conventional ads. People tend to ignore them because theyre so used to seeing them. Thats why ad companies are developing interactive and more interesting, eye catching ads. Its presumably harder on teh net though. Tbo 157 16:13, 18 December 2007 (UTC)
Exactly. On a site like this one regular users become immune to adverts very fast. But the adverts are really for the "drive by" user, who came, didn't like it, found an advert and left by that route. By all means click a truly relevant advert, but only when you do want to respond to it, not "just to see".
Getting to our "freedom target" will be several months yet. The Adsense world is a fickle world. Today, in the last 9 hours, TSW has made $1.28. In the entirety of yesterday it made $1.46. We depend on word of mouth (we don't advertise ourselves because this is not commercial exercise) and creation of good, unique articles. We're doing very well for a site launched in February 2007. But Planes is doing poorly and was launched very shortly after. Everything depends on catching the imaginmation of keen people. And it depends on making sure that those keen people have the flexibility to make it their own. We really hope clubs will start to use us as a resource. We're happy for what WP considers to be spam links (we also use rel="nofollow" to avoid being seen as a link farm - we're quite bright!) Tim Trent Talk to me 17:22, 18 December 2007 (UTC)
Ok Tim. Well, what i'll do is put around the site to some of the groups im in, informally (i.e. just word-of-mouth type stuff), and we'll see what we get. As i've said on EdJogg's page, we need to do some clearing up before we can properly launch it.
Would you possibly be able to give us an update how were doing each month ads wise? If it's not likely to be soon, may I start adding buses/trucks/cars to SW? Cheers, Bluegoblin7 18:45, 18 December 2007 (UTC)
Do we get any revenue for simply displaying the ads?
As for other content -- I think we need to concentrate on getting 'Trains' up-to-scratch, and then copying it as a working example into the other areas (such as Planes). Diluting with content regarding trucks, buses and even steam engines (though some may creep in!) is also diluting maintenance effort.
TSW is not very obvious on Google. Are you trying anything to make its presence more visible?
EdJogg 19:16, 18 December 2007 (UTC)

(reset indent)
We get a very trivial sum for simply displaying them. Adwords customers may "Site Target" adverts and say how many cents per thousand impressions they will pay. As our visitor count increases so will the fees google charges advertisers rise. Of course we never have any idea of the revenue split between us and Google. Tim Trent Talk to me 21:36, 18 December 2007 (UTC)

Perhaps aswell as the general newsletter, we should have like an admin one or something, with things like this included, e.g. the revenue for each week so that we can see how were getting on etc? BG7 21:39, 18 December 2007 (UTC)
(after edit conflict) Ah, and Google visibility. The benefit and the challenge is that we have a lot of duplicate content with WP. As content diverges our profile will rise. We are well optimised within the constraints of the software for search engines, but content is king. We submit daily sitemap updates to Google and Yahoo, and our RSS feed is registered with all the major places these things need to be registered. There is very little other SEO that can be done that is not tricksy and deprecated by Google etc. It is always posisbel to place the relevant links to us in genuine posts om message forums, but spamming is a poor idea. Tim Trent Talk to me 21:43, 18 December 2007 (UTC)

Main page question

Is it possible to hide the title at the top of the Main page like on en wp. Tbo 157 19:11, 16 December 2007 (UTC)

U mite b able 2 find out on the MediaWiki site... --Bluegoblin7 19:20, 16 December 2007 (UTC)
Well, I tried! http://www.mediawiki.org/wiki/Extension:NoTitle seems to be it. But we need to add the extension. Tim Trent Talk to me 20:29, 16 December 2007 (UTC)
Ok. Thanks. Does the extension need to added to the actual mediawiki software or is it doable without altering the software? Tbo 157 20:49, 16 December 2007 (UTC)
Hmm. wp:Special:Version does not show it in use. Feel free to ask at Wikipedia how they do it. Extensions alwasy have to be added, and we add those in the various server conmfig files. Tim Trent Talk to me 20:55, 16 December 2007 (UTC)
Also, is it ok if I put a logo of planespottingworld in the sister projects section of the main page. Thanks. Tbo 157 20:58, 16 December 2007 (UTC)
Feel free :) Tim Trent Talk to me 21:01, 16 December 2007 (UTC)
Thanks. Tbo 157 21:35, 16 December 2007 (UTC)
Wikipedia hide the title on the main page using coding in wp:MediaWiki:Common.js and wp:MediaWiki:Monobook.css. Could you take a look and try it out. Thanks. Tbo 157 12:54, 17 December 2007 (UTC)
Spoon feed me! Tim Trent Talk to me 23:47, 17 December 2007 (UTC)

Namespaces...

Hi Tim,

Looking through some archives i noticed that you needed to "create" a namespace when you created FanFic. Does the same need to be done for Workforce and Portal? No problems with Workoforce so far, but thought i should check before creating the first portal...

Cheers

Bluegoblin7 13:41, 17 December 2007 (UTC)

I didn't actually create a namespace. I created a pseudo-namespace by adopting a convention. WP does the same. Namespaces are interesting and slightly scary things. Tim Trent Talk to me 22:05, 21 December 2007 (UTC)

New bugs page

I have created Trains:Bugs and problems where users of the community can discuss bugs, possible problems and possible fixes. There is a template (in the edit section) on the page which can be used to list problems. I will also create a similar page at [2] and [3] if thats ok. Thanks. Tbo 157 17:14, 17 December 2007 (UTC)

I suggest it may be far better, since the software is identical, to create it at sworld:Bugs and problems
Never forget we interwiki link :) Tim Trent Talk to me 15:41, 18 December 2007 (UTC)
Ok. Ill move the page there and Ill put in a redirect from the one here. Thanks. Tbo 157 16:14, 18 December 2007 (UTC)

Coding followup

Ive asked for help on the reference desk on english wikipedia and the user helping has asked a few questions regarding the coding at [[4]]. So you may want to comment there. Thanks. Tbo 157 23:04, 14 December 2007 (UTC)

I have no ability to do anything today. It might be better, too, to get BG7, who created it, to comment, I think. I watched in awe as he got as far as he did :) Tim Trent Talk to me 09:18, 15 December 2007 (UTC)
Pesumably he used the skeleton portal available as a template from wp but for some reason the coding does not work here. Its already thrown 1 wikipedia user so Ill try to find someone who can help. Tbo 157 13:08, 15 December 2007 (UTC)
I do truly think the slution I put on the Cafe is the one. I just have no personal time to try it. WP is far more tolerant of bad code than we are. We also try our best to make it clean html automatically. Tim Trent Talk to me 14:33, 15 December 2007 (UTC)
Honestly, you lot, it was about 15 minutes work - comparing the WP originals and making them work here as per Tim's suggestion. Have a look at Main Page/Sandbox now... Part of the problem was that the header template was including the footer template too - the WP version used noincludes around this bit, and they had been 'lost' from the TSW version. For the footer, Tim was spot-on. We (probably, I) had deleted the odd /div and hence the page loooked OK but the template stopped working.
EdJogg 16:43, 15 December 2007 (UTC)
Way to go! I definitely needed "my" concept to be sense checked, though. I'm running on painkillers right now, so I could handle theory, but the practical application needed more focus than I have at present. To the victor go the spoils (Er, I mean would you like to move the sandbox over the current main page?") And would some clever bloke (or bloke-ette) like to see if a colour gradient is possible in box headers? And then get the colour to match the logo! Probably a CSS thing. I recommend http://www.w3schools.com/ if CSS is not your thing. We CAN create a class on our global CSS page if you make a pleasing and unique one. Tim Trent Talk to me 17:14, 15 December 2007 (UTC)
Great work Edjogg. Ill get back to working on the design now and Ill see if I can try a gradient colour scheme. Tbo 157 18:06, 15 December 2007 (UTC)
Just to let you know, Ive brushed up the layout at Main Page/Sandbox. Suggestions for further design aspects are welcome from all users. Thanks. Tbo 157 20:06, 15 December 2007 (UTC)


Bots

I have been learning what a bot "really" is. The answer is quite scary!

what it is

A bot is a combination of a user login (as the bot) and software running on a server, any server, that interacts with this place.

If you look at a bot as an intelligent keystroke macro that you execute from just anywhere with TSW as the target for its work, there you have it.

Now that does pose a few issues. The first is that we cannot open up our security on the main server to allow anyone else in. That means that we can't mount a bot for a user because we woudl also have to give some server admin rights to the bot owner in order to maintain the bot. This is why, even at Wikipedia, an external server is used. But we don't have the budget for an external server ourselves, nor will we for the foreseeable future.

So, currently, bots need to be hosted and executed from "your" own server, which makes it kind of tough if you don't have one. But you do! We all do!

Server

If the bot is to be run occasionally then the server may be the computer through which you connect to the net - your Mac or PC. Remember it is effectively a macro.

Licencing

I thought a bot was GFDL licenced. That is not the case. If the bot is written by someone and not mounted under a GFDL licenced arrangement then that bot is copyright its author. We may approach and request, but we may not simply take.

Tim Trent Talk to me 11:03, 19 December 2007 (UTC)

Can anyone find any bot experts who would be willing to set up all this. Tbo 157 11:29, 19 December 2007 (UTC)
Oddly one needs a nodding expertise with a scripting langauge, but that is all. It appears that each user can quite easily run a bot.
Now there is a different discussion to be held about who is allowed to run a bot, but that is quite a different matter.
Tim Trent Talk to me 13:44, 19 December 2007 (UTC)

Who can run bots?

Well, I would say only admins as other users may abuse them. Admins are also the only ones who can turn it off. If someone can write me some code... I'm happy to run one! BG7 13:47, 19 December 2007 (UTC)

As an admin one can turn the user id off from which the bot runs. But one cannot turn the bot off. An excellent example was the vandalbot that moved a load of our pages from Foo to Foo on wheels a while back. We blocked it for now because we have gone away from anonymous edits, but, if it comes back in an ID, which is perfectly possible, what then?
WP's bot approval (etc) policy is unenforceable. Anyone can run a bot on WP. Look how we started this place. We did not manually copy the articles. We had a bot crawl links to a certain depth and grabe the articles and their associated media (Lawfully, I hasten to add) without any WP admin rights and without any express permission to run a bot. That bot was mounted on one of Bernd's machines - I suspect it was his laptop - and ran overnight inhaling pages.
That bot could easily have been one that altered articles subtly and saved them back on WP. "Count to find the 15th "and" in an article and change it to a random conjunction from a list; wait 15 minutes and hit another."
Tim Trent Talk to me 13:58, 19 December 2007 (UTC)
Ok I get it. Couldn't Bernd code us a couple of bots?
BG7 14:05, 19 December 2007 (UTC)
He could. But it would have to interest him. I mean as a technical challenge. His day job is as head of development somewhere quite exciting, and the systems he creates are amazing and in the realms of "rather difficult and scary". His view tends to be like mine: "If we absolutely need it to make the place work, then ok, but, if it's just for fun, or to make a task easier, unless it's an interesting task, then no, thank you."
I'm sure he will be happy to offer advice, but he'd far rather teach fishing than give you a fish.
I'll ask him to look at this conversation, but he tends to avoid being drawn in. Think of him as the Keyser Söze of wikidom!
Tim Trent Talk to me 14:45, 19 December 2007 (UTC)

Should we recommend that bots run under dedicated logins? If an automated bot goes wonky, we can temporarily block the bot while the bot owner still has editing privileges to respond on the appropriate talk pages. Based on a couple questions I've received, I've been looking into how the assessment bots are run at WP (although not very deeply yet) with an eye toward setting up similar processes. Slambo 14:10, 19 December 2007 (UTC)

Yes that would probably work best - with the login being the name of the bot. BG7 14:11, 19 December 2007 (UTC)
Well, then, <posting a flag in the sand> I officially claim Slambot (talk · contribs) for this purpose. B-) Slambo 14:23, 19 December 2007 (UTC)
Ok then. If that's how were doing this: I have a friend who may be able to make me a bot to use, so i'm going to claim Goblinbot (talk · contribs)! BG7 14:25, 19 December 2007 (UTC)

Plea for good sense

Please check bots out, insofar as is possible, on sandboxes, or, a very restricted set of articles that are easy to revert. There is some form of mass reversion tool, but I have never used it, that is said to revert all edits by a named user in a named time period. I can't remember where I saw that, though.

While we take a daily incremental backup the idea of having to reciover from it scares me rigid

Tim Trent Talk to me 14:45, 19 December 2007 (UTC)

I've asked at WP about mass reversion. You mayall wish to track the answers. Tim Trent Talk to me 15:03, 19 December 2007 (UTC)

Way around IRC?

Hey all i think I have a way around the IRC problem if someone can program it. In theory, the bot can trawl through recent changes, clicking on each diff button, doing what it does, and then waiting a period of time before doing it again. Same should work for the rest of it... BG7 19:01, 19 December 2007 (UTC)

Precisely, or the RSS feed associated with recent changes which displays the diffs Tim Trent Talk to me 21:55, 21 December 2007 (UTC)

Assessment thoughts

Before we get too carried away in copying existing processes from WP, let's take a moment to think about assessments. One comment that I've frequently seen over there is that the quality assessment should be the same for all associated projects ("a stub by any other name..."), but the priorities per project could and should be different.

We can easily create a banner that will show the assessed quality, the problem is in attaching the associated workgroups to the quality in such a way that when the quality is reassessed, it updates in all of the workgroups with just one change. So the talk page would look something like:

{{Quality|stub|groups=Group1, Group2}}
 {{Group1|importance=low}}
 {{Group2|importance=high}}

What I'm not sure about yet is how to code the handling of the groups parameter in this example. It needs to be expandable to an arbitrary number of workforces so we don't run into problems when there are more than X number of workforces to list, and it needs to carry the class parameter value through to the appropriate categories so the editor only needs to update the class on one banner. Slambo (Speak) 15:25, 19 December 2007 (UTC)

Before we actually get a system in place, my comments on the rankings:
  • Quality
    • Stub
    • Start
    • C
    • B
    • A
    • FA
  • Importance
    • Top
    • High
    • Mid
    • Low
    • Unimportant
On top of these, the relevant extras, Portal, Dab etc.
BG7 15:32, 19 December 2007 (UTC)
Leaving out GA seems like a good thing here, based on the number of arguments that I see for it over there. It seems to me that we can use A in place of GA. But, do we need a C class? Slambo (Speak) 15:47, 19 December 2007 (UTC)
As I see it GA and A are duplicates of each other, so A = GA, B = B, and C = a new one, but we don't need to have it. Maybe start without C, and add it if needed. BG7 15:52, 19 December 2007 (UTC)

Bots that inspect changes

This is a challenge for us for one simple reason. We do not have the ability to use bots structured in a simlar manner to Wikipedia because theirs make use of an IRC channel to see what changes. Bots like Cluebot which [patrol and set vandalism right there are not available to us.

So I was considering mechanisms for bots to inspect in order to detect changes.

Approach 1

We have Special:Recentchanges which could be polled and inspected. And that page has http://train.spottingworld.com/wiki/index.php?title=Special:Recentchanges&feed=atom or http://train.spottingworld.com/wiki/index.php?title=Special:Recentchanges&feed=rss which update regularly.

We can't easily let bots into the data structures partly because they are weird, wonderful and fragile and partly because doing so woudl compromise system integrity.

It's a long time since I was a programmer, and I programmed in Burroughs Extended Algol, and a weird language RPL11M, but I think one might prevail upon Cluebot;s author to allow us to have a version that works (if a decent programmer modifies it) with (eg) RSS, atom or polling the page itself. If we're really nice to him he might even let us mount it on his own server

If we can achieve this (and do rememebr that Cluebot also warns vandals etc, and interfaces with WP's blocking policy which we may not want to emulate), then we may be able to reopen the door to anonymous edits.

Approach 2

write an extension that keeps a rotating logfile of recent changes which can be web-accessed, and then take it from there

Or there will be other ways. The floor is open to you guys.

Tim Trent Talk to me 14:51, 20 December 2007 (UTC)

Discussion

I've contacted another MW site about their experience of IP edits, and they get very little vandalism - if any. So why can't we just try IP editing for a week and see what happens? BG7 15:04, 20 December 2007 (UTC)

I agree with a trial. We could also work on cluebots which don't need to use IRC. There should be away to create bots which can patrol recent changes and look for characteristics of vandal edits. Tbo 157 15:08, 20 December 2007 (UTC)
Guys, please give me a break over this. The other wiki was probably not attacked yet. We came under heavy attack - we're quite visible. Pages were being vandalised faster than I coudl stop the vandalism by blocking IPs. Since we closed the door our vandalism level has dropped to zero
Get us a working anti vandal bot and I've said I will review it.
Tim Trent Talk to me 15:49, 20 December 2007 (UTC)
Im trying to find otu if tehre are any which don't use IRC. Tbo 157 15:52, 20 December 2007 (UTC)

CSS

Special:Allmessages may be of interest to admins here. It allows you to edit the MediaWiki software in just about any way. Tbo 157 13:45, 21 December 2007 (UTC)

Please only make one change at a time, test what has happened, and roll it back if at all unsure. There is scope for totally banjaxing the place here. Tim Trent Talk to me 22:04, 21 December 2007 (UTC)

Code Confilict

I noticed last night that everytime I tried to revert I was asked that mathematical question, I'd answer it correctly and then I would get the prompt that I hadn't entered an edit summary when I did. It seems to get stuck on a loop and would not let me save the edit until I turned off the prompt option. Not sure what's going on with that.

--DP67 (talk/contribs) 15:57, 26 December 2007 (UTC)
Me neither. "The prompt option"? Tim Trent Talk to me 16:15, 26 December 2007 (UTC)
Ah. User preferences. I have no idea. Tim Trent Talk to me 16:17, 26 December 2007 (UTC)
The prompt before saving without edit summary option. Although I had an edit summary it still wouldn't save it got stuck in a loop.
--DP67 (talk/contribs) 16:58, 26 December 2007 (UTC)
Looks like a bug in the CAPTCHA extension. No hope of having it solved though. Silly thought, but has rebooting your machine solved it? If so I blame M$oft! Tim Trent Talk to me 18:03, 26 December 2007 (UTC)

Editing system messages

Please be very careful when editing system messages that contain "dollar strings". an example is MediaWiki:Whitelistedittext which contains the login instructions for IP users. You will see the line:

That couldn't be easier. follow the $1 link and create your account.

If you are logged in, or even if you are an IP user and are viewing this page directly, it has the string "$1" in it. That looks wrong, but it is not. It comes into play when the page is invoked (not when it is read)

To prove this to yourself, log out, go to a random unprotected page and click the edit tab. Note how the $1 is now replaced by what ought to be there. And note how there is more than just a login action that happens when you log in by this route.

These pages are complex. Please only edit them if you know what you are doing for certain, have tested any comment you receive that they are not working correctly by invoking them, not by reading them, and by documenting what you have done on the talk page.

I only caught this one by luck. I know it was a simple error made in good faith, and I know that the documentation on the dollar strings is impossible to find.

Tim Trent Talk to me 00:50, 31 December 2007 (UTC)

Please read Mediawiki whcih describes in minimal detail a woefully small subset of the dollar string parameter passing if you dig deep enough. Tim Trent Talk to me 01:03, 31 December 2007 (UTC)
Sorry.. :( I'll keep that in mind next time. I thought it was an error.
--DP67 (talk/contribs) 02:54, 31 December 2007 (UTC)
I'm not at all worried. The great thing about wikis is that we can keep track of changes. I posted about it here so your accident, made in good faith, can show everyone else something we all need to learn. This parameter passing is a "well kept secret". Tim Trent Talk to me 09:38, 31 December 2007 (UTC)

html/JavaScript countdown timer...

Does anyone know any code for the above, that has Days, Months, Hours, Minutes, Seconds etc?

If so can you drop it in my Sandbox.

Thanks,

BG7 17:45, 3 January 2008 (UTC)

There are various codes around on the web. Search it on google. You would need to paste the primary coding into a .js page and then put the coding in for the timer on the page you want it to display. Class91 18:05, 3 January 2008 (UTC)
???... Sorry... you've kinda confuzzled me a bit there mate!
Are you trying to say search google for code, then do something with it?
BG7 18:11, 3 January 2008 (UTC)
Ye. Class91 18:37, 3 January 2008 (UTC)
Ok, I have my code, so if i put it at countdown.js what would i do next? BG7 18:39, 3 January 2008 (UTC)
I'm not sure it would work without being installed as a global script. To the best of my knowledge the $wgAllowUserJs or $wgAllowUserCss are not turned on in LocalSettings. If they were it would be nice.
--DP67 (talk/contribs) 20:26, 3 January 2008 (UTC)

MediaWiki:Watchlist-details

Yes, it is at wp:Special:Allmessages, and on WP it works. No, it is not in Special:Allmessages here, and, though the page exists, it does nothing currently.

We have no reason not to have it, but have no idea yet how to make it work. So, over you you guys. If you can tell us how to make it work then, provided it is not a MW release issue we can implement it.

Tim Trent Talk to me 00:05, 7 January 2008 (UTC)

Date links and user preferences

Is it just me or are the date links not honoring user preferences for their display? I have the "16:12, January 15, 2001" format selected in my preferences, but I don't see any difference in the date links; linked dates are displayed in whatever format they were coded in. Slambo (Speak) 16:20, 16 January 2008 (UTC)

I may have found the answer... Meta mentions that the variable $wgUseDynamicDates controls the autoformatting; if it's set to true, it autoformats date links according to user preferences. A little further digging says that this variable defaults to false. Since I don't see this variable listed on Special:Version, I'm assuming that it's unset, which would mean that the default value of false is being used here. Slambo (Speak) 20:09, 16 January 2008 (UTC)
We have set the flag. We do not yet see any changes. Please watch and see if anything happens because it may be cache. Tim Trent Talk to me 23:19, 16 January 2008 (UTC)
Hmmm. Bernd says it does work but only on history pages. Ok, that is just weird. Tim Trent Talk to me 23:23, 16 January 2008 (UTC)
Now it appears to be working for me. Thanks! Slambo (Speak) 12:24, 17 January 2008 (UTC)

Possible MW extensions?

Hello all,

Just wondering if anyone thought that installing the following MW extensions would be a good idea?

(links to follow shortly!)

  1. Dismissable Site Notice
  2. Giverollback
  3. Oversight (as we have young users they may post stuff not suitable?!?!)

Thanks,

BG7 10:15, 2 April 2008 (UTC)

Permanent Message on the anon site notice?

Hello all,

Just wondering if we should add a permanent message on the anon site notice saying "Please log in" or similar?

Thanks,

BG7 10:16, 2 April 2008 (UTC)

We have server constraints

Our host has constrained our server. We expect this to be resolved by negotiation or migration within 2 weeks. Tim Trent Talk to me 09:58, 23 May 2008 (UTC)

Server migration is completed

Much planning, but now we are on a bigger, faster server. Tim Trent Talk to me 23:35, 7 June 2008 (UTC)

Good stuff! Now we can make a larger mess more quickly!! :o) EdJogg 10:44, 9 June 2008 (UTC)

Performance enhancement

We have now migrated all SpottingWorld to a substantially faster web environment on the new server.

This ought to make life even easier.

Tim Trent Talk to me 15:54, 22 June 2008 (UTC)

Navbox upgrade issues

If you have a look at today's recent changes you will see a surge of template activity. This was precipitated by someone making a real edit (shock horror) to Manchester Metrolink, and me spotting that the page was missing several templates.

As part of this process I needed to fetch {{Navbox subgroup}} and, to cut a long story short, realised that the main {{Navbox}} template here was very out-of-date compared to the Wikipedia version on which {{Navbox subgroup}} was based. To cut another long story short, an attempt to update the Navbox template (to allow import of new naviboxes) failed miserably. The problem may be a missing CSS style, but I am no expert on these so I have made no further progress. I have reverted the changes and if someone would like to try resolving the matter, we may be able to re-enable the new versions.

NOTE - since the Navbox is the basis for hundreds of other templates, use of a sandbox is encouraged when trying to make it work!

EdJogg 16:58, 28 October 2008 (UTC)