Wikinews:Water cooler/technical/Archive/3

News ticker

A small, unintrusive news ticker could be added to the template on the front page. It would get the headlines dynamically from all the latest news stories.

It can be done with:

<marquee bgcolor="#CCCCCC" scrollamount="7" scrolldelay="0" height="30">(This part is retrieved from the latest headlines)</marquee>

(above comment posted by anonymous contributor)

I've never seen such a thing on a Wiki site. Is this supported by the MediaWiki software? I don't think you can add arbitrary HTML to a page, just the small subset of what is supported in MediaWiki. — DV 10:30, 28 Feb 2005 (UTC)

Oh, it's probably quite possible, but that doesn't mean it should be done. Marquee is a nonstandard tag and is very uncommon nowadays, it's use should be avoided. I'm no modern HTML expert, but I must think that there would be a better way to do this.--24.70.95.205 00:42, 2 Mar 2005 (UTC)

It can be done with JavaScript - various examples can be found on google. Java can be used too, or flash, both of which have the added advantage of being able to easily retrieve up-to-date news, even if the page isn't refreshed. EAi 02:04, 2 Mar 2005 (UTC)

isn't the marquee tag IE only? 24.65.22.111 01:27, 9 Mar 2005 (UTC)

Works fine in Safari. Ludraman 23:55, 10 Mar 2005 (UTC)

Ack - please no scrolls or popups or animations or .wav files -- Davodd | Talk 18:09, 15 Mar 2005 (UTC)

Just a thought: Perhaps we can have some kind of ticker generated automatically with our latest headlines for off-site use, which should generate traffic back here. Same can be said for the RSS feed, but there are still a lot of issues with that imho. -- Redge (Talk) 09:23, 13 Apr 2005 (UTC)

Separate templates on Editor's tasks

Just wondering - why? It adds another layer of complexity to an already very complex site Dan100 (Talk) 09:51, 2 Mar 2005 (UTC)

I broke Editor's tasks into separate templates. My main motivation was so that I could get rid of the section edit links. The problem was that since Developing articles was already a template, the edit link didn't let you add stories-- just add or remove the template. To fix this, someone added a link that directed you to the editing page of the template. That worked, but then we have two links that say "edit," with no distinction between them, and only one worked. That confuses people, which is bad.

Given that, I wanted to either make them all into templates, or none of them. I chose the more complex route, but I wouldn't mind if we decided to make it ALL one template-- I just want the edit links to be obvious to new people. I figured that as the site grows, we'll start running into edit conflicts if we keep it all as one page.

One other idea-- maybe we could make the smaller, rarely used sections of Editor's tasks back into a single file, and fix the section links by some other means. That might be the best solution, and I think I know how to do it-- maybe tonight. Pingswept 14:41, 2 Mar 2005 (UTC)

Good reasoning. I hadn't thought about the potential for edit conflicts. I think we should stick to what you've set up Dan100 (Talk) 17:55, 3 Mar 2005 (UTC)

Stock market data - help needed

Folks, I want to put together a bot (using the Wikinewsbot framework) to automatically update Wikinews with financial market information. However, I am unable to find a good source for this data. In addition I don't even know if:

it is legal to screen-scrape this from somewhere (after all, this is factual data which is not copyrightable)?
- what about reading CSV data that's provided for download by a service like Yahoo?
is there a free web service of some sort for this data, worldwide?
is there even a place other than Yahoo Finance that has a list of international market data?

Let me know if you have ideas. Otherwise I'll probably write a quick bot to read in Yahoo's CSV files, which seems legal to me, and post them here once a day. -- IlyaHaykinson 03:54, 5 Mar 2005 (UTC)

Alright, so since nobody responded in time, I've gone ahead and implemented this. It's in testing at my wiki: see http://www.ilya.us/wiki/index.php?title=Market_Data for a list that auto-updated every 6 hours (for now).

I'm also thinking that we could have the bot create and maintain Template:Market_Data_INDEXNAME, so that editors of local regions could just {{include}} a custom set of templates in. What do you folks think? -- IlyaHaykinson 10:52, 5 Mar 2005 (UTC)

Brillent Ilya! I've been (quite irregularly) manually updating the figures using BBC Markets for the past few weeks. Your bot makes it much easier. Im not certain about the legality of copying the date, many stockmarkets [1], [2] have there own figures listed on there websites. So there providing the figures free of charge. One thing to remember, all quotes on all media sites (yahoo, bbc, marketwatch etc.) are 15/20min delayed quotes. Thanks for your excellent work. CGorman 15:15, 5 Mar 2005 (UTC)

I am not a lawyer (but you probably already knew that, otherwise I wouldn't be offering my opinion for free) but there is something called a "misappropriation of hot news" that is an exemption to the general ability to scrape data from other sites. This should be reviewed against Wikinews offerings by those qualified to do so. See for example http://cyber.law.harvard.edu/metaschool/fisher/linking/doctrine/. This applies particularly to scooping real-time data (I think). DouglasGreen 17:28, 25 Apr 2005 (UTC)

Stock market data - help offered

My wife and I both enjoy watching the stock market (it's the first thing she checks when she wakes up each day), so I'm happy as a peach that you've chosen to pursue this project.

Although I'll continue to focus primarily on WeatherChecker until it is more capable, if you make the source available for your "MarketBot", I would enjoy helping to develop an application to generate market chart images, which would do a better job of visually communicating the market data over time.

We could call this new application "MarketChecker".

I'd also like to expand the coverage to include widely-held stocks, (starting with the Dow components).

We could then have an easy to understand Markets section on the Business page, with thumbnails for various charts, and links that would lead to expanded coverage on a "Markets" sub-page. Once this page was ready, we could then put a small version of the daily chart for two or three major market indices (updated hourly during market hours) on a new section of the Main Page, as an enticement to visit the Business and Markets pages.

At a minimum, I'd like to see an up or down arrow next to the current DJIA, NYSE and NASDAQ market values, indicating the recent trend, somewhere on the front page.

Please let me know when you're ready to share the "MarketBot" source, and I will put together some sample graphics to illustrate some of these ideas.

— DV 15:59, 5 Mar 2005 (UTC)

P.S.: Ilya, I sent you an e-mail with a postscript to these remarks. — DV 16:20, 5 Mar 2005 (UTC)

I've checked in what I'm running now into the Wikinewsbot project. It's called the MarketDataBot for now :)

Here's what I think it should do, in the short term (assume ^DJI stands for any of the tracked indexes)

Post updated summaries to [[Template:Market Data/^DJI/Latest]] using some {{index-data-summary}} template, somewhat like it's done on Ilyawiki now but with a separate page for each index
- This would let people pick and choose which indexes they want to show on other pages.
Post the historical data as an XML file into [[Template:Market Data/^DJI/Data/Historical]]
- The bot would retrieve the data from there to generate charts, later
Post an updated [[Template:Market Data]] which includes all of the updated per-index templates

Manually, we would maintain:

[[Market Data]] would include {{Market Data}} and thus get all of the necessary indexes
[[Market Data/^DJI]] would include more information about the index and would later include any charts

The problem I generally see with generating charts is that we don't have historical information for these indexes now, and won't have it till we run the bot for a while. I would thus urge us to run this type of a bot for a bit before getting into charting. We also don't have a way to programmatically determining components of an index.

Oh. And we should get an account used to do this marked as a bot — anyone know how to do this? This way its changes won't clutter RC. -- IlyaHaykinson 22:09, 5 Mar 2005 (UTC)

Can't we populate the history by just downloading the data and converting it into our own log file format? For example, here is the history for ^DJI ready to use in spreadsheet form.

Now that WeatherChecker is updating hourly, why not do the same for "MarketChecker" and generate hourly charts showing the progress of each index from the opening bell?

At the close of trading, MarketChecker could generate a daily chart by compositing the day's data with the pre-populated history.

We won't have hourly data for previous days, but our multi-day charts can just show closing prices for days previous to when we started. The effect will be that the chart will be coarse for all of days prior to our service becoming operational, but then the chart will appear very fine-grained starting from the day the service has launched.

I'll take a look at MarketDataBot, and then determine what would be required to pre-populate our history and begin the MarketChecker charting application.

Market reporting without commercials - now there's a concept! I bet a lot of folks are tired of being ambushed by commercials for low-price stock trades every time they lookup market news - so this will be a very attractive service for users who just want to concentrate on the market without distractions.

— DV 22:32, 5 Mar 2005 (UTC)

By the way, why would we need to programmatically determine the components of an index? The value of all major indices is quoted, so it's not as if we would need to calculate the value of an index by referring to its components.

However, it would be useful to list individual quotes and charts for the Dow 30 components, as well as other widely-held stocks. — DV 22:47, 5 Mar 2005 (UTC)

I've gone ahead and implemented changes roughly how I outlined them earlier, complete with a bit of charting. Check them out at http://www.ilya.us/wiki/index.php?title=Market_Data and http://www.ilya.us/wiki/index.php?title=Market_Data/%5eDJI for one of the indexes, specifically.

Thanks DV for the pointers to the historical data on Yahoo.

There's no intra-day charting (and probably we should test in this mode first, before moving on to more / more-frequent updates (the data on ilyawiki is updated every 6 hours, currently). -- IlyaHaykinson 11:51, 6 Mar 2005 (UTC)

Yahoo Finance UK [3] has even more world stock indices than the Yahoo US website.

Limiting to be useful

Although I am for inclusievness, maybe we should have a separate, Major Indices section - since not all stock markets have equal weight on the world economy (or may not be of much interest to an English-speaking audience). I suggest an idex with the following: All Ordinaries (Australia), CAC 40 (France), Dow Jones Industrials (USA), DAX (Germany), FTSE 100 (United Kingdom), Hang Seng (Hong Kong), Nikkei 225 (Japan), Straits Times (Singapore), Tronoto Stock Exchange (Canada). -- Davodd | Talk 18:48, 22 Mar 2005 (UTC)

It looks to me as if commodity markets are springing into the consciousness of seasoned stock investors(aka gamblers). With various hedging strategies they are no riskier than stocks on margin. Even a market for shorting house prices is being developed.. so it may be useful AND cutting edge to include the majors' (London/CBOT/NYMEX)prices on currencies at least. Paulrevere2005 21:39, 23 Mar 2005 (UTC)

I will launch the bot on Wikinews tonight. There will be one main page, Market Data, which will have all of the indices that the bot will track. Each index is updated via a template like Template:Market Data/^DJI/Latest and the image will be included via another template, so each region, and the homepage, can pick and choose what they display. In addition each index's landing page (Market Data/^DJI as an example) is manually maintained to contain whatever info one wants to have there. I will update here when the bot starts running.

As an aside: Yahoo's historical volume information seems to have gotten rather spotty. If someone can find a better source for this data, please let me know. Remember that it needs to be international and somewhat centralized, and computer-readable. If interest is high enough, we can apply for the Wikimedia Foundation to pay for access to this data from a commercial provider, but that's a bit iffy. -- IlyaHaykinson 22:36, 23 Mar 2005 (UTC)

Alright, it's launched. A selection of indices is on the homepage. The full list, maintained every 6 hours, is at Market Data. The templates are modified by the bot: other pages can be human-maintained. IlyaHaykinson 01:27, 24 Mar 2005 (UTC)

Dutch interface links

I had Dutch interface activated in my preferences, but I quickly found the Dutch interface link description don't match their targets. For instance, clicking "In het Nieuws" (In the news) brought me to Commons upload. I'd be more than willing to fix this, but I'm not sure where the Dutch interface system messages are located. Anyone know where that would be? -- Redge (Talk) 12:39, 17 Mar 2005 (UTC)

Date Tag

After having a discussion with Amgine about the date tag, I've heard there are no links from an article of a certain date to a news archive. Perhaps we could alter the date tag to provide a link to the archive as Amgine has done in User:Amgine/Sandbox6? -- Redge (Talk) 16:48, 21 Mar 2005 (UTC)

I don't see why we have to hard-code the year, nor link to those archives separately, when they're already linked to from the day's archive page. What about something more like User:IlyaHaykinson/DateSandbox (and the test at User:IlyaHaykinson/DateSandboxTest)? -- IlyaHaykinson 18:48, 21 Mar 2005 (UTC)
Why do we need templates to do this at all? According to Help:Editing#Links.2C_URLs there's a feature of the MediaWiki software that will both automatically turn a wikified date into a hyperlink and display the date according to each individual user's preferences. Can we not get that feature enabled and have it configured as appropriate for the WikiNews naming conventions? (Or is the problem that it isn't configurable?) It would be much easier to just enter wikified dates, as the editing help describes, and have done with it. Uncle G 20:21, 30 Mar 2005 (UTC)

Bundling articles

It seems to me that, as we have lots of articles coming in that are related, like a string of articles about the current Ukrainian situation, we should find some way of simply linking those articles, bundling them in a way. A few example methods of the top of my head:

Create a master article with a list of articles relevant to the topic
Create some kind of custom category for each topic
Create some kind of custom template for each topic

Problem with the first two is that you don't have direct links in the articles themselves, the problem with the latter two is a surplus of categories and templates. I hope someone has a better idea? -- Redge (Talk) 20:38, 22 Mar 2005 (UTC)

All three can, and have, been done here. Categories are almost a given; each of the Ukrainian documents, for example, should have the [[Category:Ukraine]] at the very least. A topic page would be a good idea for someone with a strong interest in the issue, perhaps similar to the 2004 Indian Ocean Tsunami in concept (but of course it would need to be designed to fit the topic. A combination of both a category and a topic page is also possible, simply building the former at the top of the latter (but realizing the category listing will be at the bottom of the article, and will be alphabetical.)

There is an "Issues" template series, though I'm unfamiliar with it. Check any of the recent North Korean articles dealing with nuclear issues for an example. Other "issues" templates have also been defined, and I think they're available via the template listings - Amgine 21:59, 23 Mar 2005 (UTC)

I would like to propose we use the recent spree of Wendy's articles to create a more effective method and use this technique consistently. What I would like to propose is create a sidebar template which can be added to the Wendy's articles. This sidebar should include:

A picture
A list of Wendy's related articles (either manually or by DPL)
A short description of the articles (as to explain the circurcumstances when the bar is placed in a new article)
The tag [[category:Wendy's]] so that all articles with the sidebar are also added to the right category
A link to the Wendy's article in the main namespace.

When the scandal has taken it's course, we can summarize the events in all the articles we have accumulated in a new article or within the Wendy's main namespace article. Any comments/suggestions/critique? -- Redge (Talk) 11:19, 16 Apr 2005 (UTC)

Axe language selection?

I think the language selection page is a little bloated to be the first page people see when they visit Wikinews. It'd be really easy to give users a PHP cookie when they select a language and redirect based on that. Or even better, redirect first-time users to the English page with a link to select a different language.

Another option is to automatically detect the user's browser language or location (almost all major news sites do it) and point the user to the appropriate page. (This is what I'm leaning towards. I could actually code this if you needed me to)

I wouldn't recommend this for Wikipedia, but because of the nature of Wikinews, I feel it's warranted. --Noclip 15:05, 23 Mar 2005 (UTC)

I'm assuming you're refering to the Wikinews Portal, which is the project's "front door". There is currently a design contest (with prizes) to change the look of that page. Most wikimedia projects have a portal page; for example Wikipedia and Wikisource, though they deal with multiple languages very differently.

Most people bookmark the domain of their language (for example, http://en.wikinews.org) rather than the main portal. - Amgine 22:09, 23 Mar 2005 (UTC)

Why direct first-time users to the English page? After all, English speakers are rather in the minority of the world's population... Dan100 (Talk) 09:30, 29 Mar 2005 (UTC)

Wikinews-specific skin (Full details)

(Please read the full details here!)

Noclip 19:06, 25 Mar 2005 (UTC)

OK Alpha geek challenge... overriding the image click link

Is there a way in wiki markup [[image:image.jpg|thumb|etc...]] to override the click-through link so that it doesn't have to go to the image page? This would be useful on the front page, latest news, and daily templates to allow a person who clicks on the picture to go to the news story - instead of always going to the wikicommons image page. -- Davodd | Talk 11:37, 28 Mar 2005 (UTC)

No, it's not possible. There is a bug requesting this feature in MediaWiki, and there has been some debate about it on Wikipedia, but right now you can't do it. Dan100 (Talk) 09:37, 29 Mar 2005 (UTC)

Davodd, this is quite possible once it is implemented in the MediaWiki software. However, relying on a "bug" in Bugzilla isn't going to get it done. Perhaps the developers have better things to do with their time than add features just for Wikinews. Or perhaps they are not interested. Wikipedia is the 800-pound gorilla after all.

So one of the "third-party" developers here on Wikinews is going to have to make it happen! So let's get started, shall we?

I've been asking for this feature for a while, but let's fully justify it in terms that appeal to me as a developer. :) In addition to the reasons you cite, we also need this feature for the weather map thumbnails, to take the browser to the associated Weather page instead of to the image file. OK, now I am motivated!

I propose this syntax:

[[image:foo.jpg|redirect:[[foo article]]|caption text...]]

One of the issues is attribution, because the viewer would not see it on Commons anymore.

Two proposals to handle that include:

Composite the attribution sideways along the edge of the image.
Show the attribution along the top or bottom of the page to which you are redirected.

There have been ongoing discussions about how to best move forward on jumping in and helping with features (Ilya and Amgine have bug 1411 in mind), but it's not as straightforward as simply opening the parser source code, making the change, and e-mailing it to Brion. :)

I plan to set up a development MediaWiki server soon, partly to simulate features under heavy load conditions. Hopefully if we can prove a given implementation will scale well, the developers will take us more seriously.

It's a bit of a chicken and egg problem. It's hard to justify simply running off and writing code to implement the features we want, without risking wasting our time because the devs think we should use a completely different approach. But they won't even take us seriously enough to discuss the proper implementation until we have some code to evaluate.

As a result, so far I've stuck with writing client-side apps.

Once I have a MediaWiki server set up, I can experiment and find out for myself what the issues really are. Various kind souls (you know who you are) have been very helpful in guiding me towards learning how to set up a MediaWiki server, which has been a big help, as getting started is half the battle.

There are those who feel we need to pay money to hire "professionals" to do the job right, because supposedly volunteers have no hope of getting anything done. Looking at Linux, I would dispute that. Who's to say a "professional" wouldn't volunteer his time for free if there was more flexibility to allow participation by new faces?

Hopefully it will not become necessary to fork the MediaWiki software, (and Wikinews itself) to add the features to the software that are required to scale up to more users.

I will post notes on my user page as this feature progresses.

— DV 16:45, 29 Mar 2005 (UTC)

As I understand it, the developers are very cautious about adding any new features to MediaWiki as the code base is pretty big and quite messy. It's pretty easy for a new feature to hurt performance, which is a big deal for something that powers Wikipedia. The developers are also busy concentrating on performance issues right now to try and keep abreast of the crazy growth Wikipedia is seeing. So I wouldn't hold out too much hope for getting a patch committed. Dan100 (Talk) 00:16, 31 Mar 2005 (UTC)

DynamicPageList

Refresh to check your tests of the below. - Amgine

Folks, thanks to brion, our case 1411 — the DynamicPageList — has been installed on Wikinews. It's in test mode, so it could get yanked out, but this is our chance to test its usefulness / usability. See the source for this section to see what the usage is.

A dynamic page list is basically a listing of links from one or more categories. The easiest example is the following, which lists all pages in Category:March 29, 2005:

You could also combine categories (i.e. returning pages in cat1 AND cat2):

Canadian Football League to test Atlantic expansion with exhibition game

finally, you can also limit the count:

If you don't want the bullets, do:

Stories are listed from most recently modified to least recently modified. You must specify at least one category, and at most three. The count cannot be above 50.

We might want to implement this on the homepage, via a special Category:Published which we could add to stories that we want to automagically appear on the homepage. Combining with such a category would then require just a cache purge of pages that should list the story.

Thanks to Amgine, wrote this extension's code with me. -- IlyaHaykinson 09:44, 30 Mar 2005 (UTC)

I was once told that this listed articles on the basis of 'most recently edited'. Is that the case? Dan100 (Talk) 00:26, 31 Mar 2005 (UTC)

It should find the most current timestamps of articles, and sort them in that order. Remember it will only find articles which match *all* the categories. (which is why we're considering creating a Category:Published, for articles which are ready for the main page.) - Amgine 00:42, 31 Mar 2005 (UTC)

What bothers me is that it would alter our current 'most recent on top' ordering of articles. This would make it quite hard for readers to see when a new story has been added, rather than an older one floating back to the top because it's been edited. However I believe sorting for the date category will stop articles from different days getting jumbled up... right? Dan100 (Talk) 00:48, 31 Mar 2005 (UTC)

<nod> Actually, it sort of gives a more accurate view of "most recent on top", when referring to edits as opposed to publishing. Yes, we will need to figure out a process to use the most current date as a separate category. This may get sticky, now that I think about it. - Amgine 00:53, 31 Mar 2005 (UTC)

Any chance it could go by major edits only? Listing by most recent edit could actually be rather good as our articles tend to develop a lot over their first few hours of life and it would serve as a way to highlight to returning readers that changes have been made. However articles going back to the top because of typo fixing etc wouldn't be so great... Dan100 (Talk) 00:57, 31 Mar 2005 (UTC)

That's an interesting idea. I think the SQL could do that, in the more current version, but it may miss articles which have not been edited in a while. So, when searching for a category which only rarely has new articles (such as obituaries), articles may not be found simply because they are older than the recentchanges table keeps track of. - Amgine 01:06, 31 Mar 2005 (UTC)

Obviously we need to test this out. But can I suggest we try it on a sandbox Main Page rather than the real one? We're getting serious traffic now and I don't think we want to risk annoying readers. However I'm keen to see this 'in action', I suggest categorising articles as 'published' when they get listed on Latest news (along with appropiate other cats) for the next 24hrs or so and seeing how it pans out. Dan100 (Talk) 00:53, 31 Mar 2005 (UTC)

As a test, I've updated the unused South America regional news page to use the dpl. - Amgine 01:06, 31 Mar 2005 (UTC)

Refresh to check your tests. - Amgine

New real-time online collaboration tool

Have fun kids! Very sweet. Firefox/other Gecko browsers only, mind you. Dan100 (Talk) 00:12, 31 Mar 2005 (UTC)

Tried it. Very slow and not really usable yet (editing the XHTML didn't work at all for me, the text document did, but with many internal server errors), but shows promise.--Eloquence 02:58, 31 Mar 2005 (UTC)

Why we need to use 'Subst:' in templates

After a recent vandalism of the source template (resulting in the vandalism of all articles on the site) I re-visisted w:Wikipedia:Avoid using meta-templates to investigate what could be done about it. One solution would be protecting the templates concerned (source and date). That wouldn't be a big deal as they are effectively static; neither has been altered for months. However it isn't very 'wiki-wayesque'. (I don't think the database load is is big deal because we're so small and naturally the number of articles only grow slowly (not like the 1,000s of new articles seen everyday on Wikipedia.

DV objects to this on the basis that it obscures the use of templates from new editors (I believe all new editors use the Workspace and associated pages so disagree), and that it precludes making site-wide changes to the affected templates (but as I said before, no-one wants to do that and "maybe one day" seems a poor reason). He also mentions copying and pasting existing uses of sources and editing sources being made harder.

Note that I'm only proposing changing the example wiki mark-up in Writing an article and Reporter's tools. Naturally I'm using Subst in my own articles.

Please read the above-linked Wikipedia article and discuss! Dan100 (Talk) 10:17, 3 Apr 2005 (UTC)

While I appreciate Dan100's attempt to paraphrase what I wrote on this subject up above, he's glossed over some important points. Rather than copy and paste a bunch of text onto this page, I ask that anyone who is interested in this topic please take a moment to read a more detailed discussion over on the Reporter's tools talk page. — DV 10:58, 3 Apr 2005 (UTC)

I've made my case above. Can I invite other editors to consider this issue? Some third-party input would be great. Dan100 (Talk) 12:37, 4 Apr 2005 (UTC)

I am ambivalent at best. Although I can see a drawback to dropping templates if categories ever change or expand. -- Davodd | Talk 01:10, 13 Apr 2005 (UTC)

How does Index work?

How does the Index work? I am referring to the box in the main paige (the one which refear to Template:FrontPageSubjectMenu. How does the subject nd geographics area work? Thas an article appear here automatically? 82.52.78.173 13:10, 5 Apr 2005 (UTC)

As of yesterday, yes articles will appear on the relevant region or topic pages automatically if editors use use the right categories (ie those on template:FrontPageSubjectMenu). Dan100 (Talk) 08:25, 7 Apr 2005 (UTC)

Since I am working in the starting up of Italian Wikines (Wikinotizie) could you tell me where can I found information on DynamicPageList? AnyFile 10:13, 10 Apr 2005 (UTC)

Minor bug in the software, or bad wiki markup?

There appears to be a minor bug in the wiki software on this site, or else the wiki markup is formed badly in a way I can't see. I have seen it eat the beginning of several articles.

Most recently, it happened to Afghan coalition forces repel Taliban ambush: 8 killed, 16 captured. When I first came to the article, the first sentence as it was rendered, read "Afghanistan Monday, 8 Taliban rebels were killed and 16 captured when Afghan coalition forces called in U.S. air power for support."

That didn't really make sense. When I looked in the source code, I found:

[[image:Zabul Province.png|thumb|right|180px|US-led coalition has 18,000 troops in Afghanistan<br> [http://en.wikipedia.org/wiki/Wikipedia:Text_of_the_GNU_Free_Documentation_License GNU Map]]]

In a convoy ambush in southeastern [[w:Afghanistan|Afghanistan]] Monday, 8 [[w:Taliban|Taliban]] rebels were killed and 16 captured when Afghan coalition forces called in [[w:United States|U.S.]] air power for support.

Notice that the introductory text, "In a convoy ambush in southeastern" was not showing in the rendered article, though it was there in the source. All the text up to the first wiki link in the article was not being displayed. I fixed it by adding a space after the GNU Map link, like this: "GNU Map] ]]". Then the text started appearing.

Just wanted to report this, thanks. DouglasGreen 23:49, 19 Apr 2005 (UTC)