My biggest pet peeve on running this blog

I understand.
You might consider if there is some code that can be automatically inserted at the end of every post that will turn off any such markup that the post has left on.
REPLY: And that is what I asked wordpress.com for…no can do -A

>>i>I’ve never had that problem. Myself, I’m not that familiar with html.
[Reply: I can tell. You did the backward “hosed” tag that Anthony wrote about, and every post after yours was in italics. Fixed now. ~dbs, mod.]

I am an old fashion kinda a guy–*stuf* or _empfassis_ is good enough for me. If you can’t do it with a 4-row teletype, it doesn’t need to be did.
What if everybody _started_ their comments with **?

@Larry Sheldon: We would be baudoted to death?

Anthony,
Many thanks for the illustration on the error of my HTML ways
on screwing up the “slash” in front of the “i” when trying to italicize.
I’m still adjusting to my new trifocals.
Don’t I recall you saying some time ago that wordpress.com isn’t
able to provide a segregated sandbox for HTML test runs on our
comments before we enter them onto the thread ?
My old Word editor won’t let me run HTML tests at home.
[Reply: use the “Test” page in the masthead menu. ~dbs]

Slash-eye

Yup global warming.

Not only every following comment, but also all the text in the sidebar.
[You da man!☺ ~dbs]

The only problem with publicising something like this is that there are people out there who will now do this sort of thing just to cause mischief.

This is really a bug with WordPress. They should automatically self-contain each comment and run a filter to weed out errant html for allowed tags. I realize this is nothing AW can fix, it’s something WP needs to fix.

WordPress could fix this iff they ran stuff through HTMLtidy before sending it to the database. I used it to fix some tragically screwed up HTML entered idiots who were incorrectly trained in how to enter HTML in our database and who insisted, even after I corrected the training advice, in entering tragically bad HTML. Tell the geeks at WordPress that HTMLtidy can fix lots of egregious problems in user entered HTML. There is no excuse.

My experience is that even knowing how to properly use HTML tags and being a veteran at it does not prevent a simple mistake in the heat of commenting.
When drafting a comment I find that using a full capability word processing program with spell checking at least brings attention to the HTML tags as being a misspelled word. So you can at least look to see if it is correctly constructed. When everything is OK in the word processor then cut and paste to the WUWT comment box.
John

How about adding “</i>” and “</b>” to the end of every post by adding it to the template code? Looking at the code, a posting on this page is always enclosed in <div class=”comment-body”> and </div>. Adding </i></b> before the </div> in the template will stop the error from mangling all other posts and unnecessary closing tags are ignored by the browser.

I will be more careful about that.
Test Bold Normal
Bob Diaz

Probably should just fix the software. There’s no excuse for the software being so bad that a formatting error in a comment breaks a page.

If you have a site admin, or have a trusted person who has at least dabbled in php, here’s what you need to do:
tell them to find the region in the commenting script that parses the database record row by row. There should be a specific line that handles the body of the comment. From there, all you have to do is something like the following:

if(stripos($database_row['comment'], '' && !stripos($database_row['comment'], '')){
$database_row['comment'] .= '';
}

where “$database_row[‘comment’]” represents the content of a database result’s comment field. This scriptlet will determine whether or not a comment has the tag without its closing tag.
You can email me at dcdorset ~@~ gmail DOT com if you need more help. I really enjoy your site, keep up the good work!

May I also suggest that you make your guide a little more explicit under where you enter your post as every-one may not know what all your HTML tags mean and how to use them:
Something like:
http://www.ibdguy.com/
for the tags you allow.
Or by adding a ‘Complete Idiots’ type guide to the RHS side of the page where you have other references.
E.G
http://dynamicdesignz.net/guides/Tags/index.htm

A Tech resolution will beat out an appeal to ignorance resolution every time. You will be frustrated forever until you outsmart the problem.

High up in the sidebar I’d put:
Close tags correctly. Right: </b> Wrong: <b/> Right: </i> Wrong: <i/>
I wish I were acquainted with WordPress. I have trouble believing that you couldn’t manually put extra closing tags into the template. I just checked and found places where it would work in Google Blogger templates, both the “classic” template and the recent one. Seems simple. (Of course, Blogger interrupts the posting of a comment with an unclosed (or badly closed) tag anyway, so I can’t test the effectiveness of adding of extra closing tags into the template.) (I’m not saying you should switch to Blogger, I know that WordPress has advantages.)

North of 43 and south of 44 said @ March 19, 2011 at 5:01 pm
“Yup global warming.”
Nope. HTML’s screwed @43degrees south and we ain’t got no global warming! If you have any spare you can send down here, it would be very much appreciated 😉

Why not just let the trolls know how to bring down your blog while you’re at it!
[Reply: it’s easy enough to fix, and if someone does it more than a couple of times, they might end up in the spam bucket permanently. ~dbs, mod.]

The problem is due to lazy coding on the part of WP, and a somewhat complicated muddle of syntax between different markup languages.
WP do in fact filter comments for unbalanced tags and other problems. That’s why attempts to fix it by other posters putting in closing tags don’t work. The filters spot these as unbalanced tags and delete them.
I think the bug is in a function called force_balance_tags. First they find a tag. If it starts with a / they assume it’s a close tag. If it doesn’t, it’s an open tag, but if an open tag ends with a /, they immediately (and incorrectly) assume that it is an XHTML self-closing tag and ignore it. They don’t check to make sure it is one of the valid self-closing tags. This means that the i/ combination gets treated as a self-closing tag and is left in. The original i-tag is closed off, but the second i/-tag is not.
Since the browser thinks it’s reading HTML, and self-closing tags are not valid in that language, the closing / gets ignored and it acts just like a normal i-tag and turns everything to italics.
I am sure it is fixable, but could be tricky to do without breaking something else, and high-level computer geeks of the sort who dream in regular expressions put a low priority on proofing things against idiots who can’t even manage to code basic html properly. They’re like that.
As I’m not such an expert, I could be completely wrong – but that’s what it looks like to me.

It would require a sea change, but the folks over at Stack Overflow (http://www.stackoverflow.com/) and friends have been using Markdown (http://daringfireball.net/projects/markdown/) rather than HTML for text formatting with great success. It started out as a programming site, but has grown a bunch of companion sites covering a wide array of technical (and even non-technical) topics, and the learning curve for the typical user seems pretty short and easy. I would wonder if WP could do that too. Markdown uses plain ascii text in ways that are familiar to typewriter (and teletype) users.

That’s why I don’t mess around with those things. I know its just asking for trouble.

What we need is a preview function.
Lots of sites have them.

The way I do it is to read the HTML comment in to an XML document, and then write out the XML. This closes unclosed tags automatically.
http://www.hoboes.com/Mimsy/hacks/auto-closing-html-tags-comments/
This could be integrated into a WordPress plugin, but I don’t know how much control WordPress gives you over custom plugins.

Anthony,
I’m a moderator on another blog, and WordPress solves this problem in the edit mode, where each comment has an HTML tool bar directly above the comment. I’ve always wondered why they couldn’t do an abbreviated version of that tool bar in the comment mode with the codes you show below the comments here. Then all the commenters would have to do is highlight the portion to be altered and click on “bold” or “italics” or “delete” or “link”, etc., much like you do in Microsoft Word.

Nullius,
You are completely wrong. HTML does have valid self closing tags, button, input, hr, br are a few of them. The problem is that HTML is an ugly language which is ill defined. There are far too many tags in HTML which do not require a closing tag. You can’t really deal with this in regular expressions and need a full parser. This is why I recommend HTMLtidy which is a full parser and detects silly mistakes and corrects them. XHTML requires closing tags (whether self closing with a / before the closing > or a / [tag] > for the closing tag. In XHTML every tag requires a closing tag. In HTML this is not the case. HTMLtidy fixes this as best it can and does a pretty good job. Once the input, as filtered by HTMLtidy is clean you can then pass it to the validator (however you implement it) and get far better results in filtering for dangerous tags.
In the system which I write and maintain that means running the input through HTMLtidy and then through an MS security library to get a safe HTML fragment for storing in the DB. There are other libraries which do the same sanitizing of HTML input for PHP. I assume that WP does sanitizing, but it is clear that they do not do correction (provided by HTMLtidy). Hey, it works for me!

So install an RTE.
So simple. Use an RTE or fix coding mistakes. As you can see from the previous comments, most people don’t know how to code. And why should they have to?
http://tinymce.moxiecode.com/

OK boys and girls.
Here is a short tutorial for all you folks with really fat fingers.
http://www.gojackets.com/help/tutorial.shtml

REPLY: And that is what I asked wordpress.com for…no can do -A
Wow, really? They said no to that? That’s terrifically bad customer service.

Maybe overly simplistic but why not just switch your commenting engine to IntenseDebate or Disqus instead of the WP default? They even import the old WP comments?

Amazing to me that WordPress is unable to resolve this very simple problem. A real simple piece of code could fix this issue easily. I use this technique in many applications that I have written. This is one of those things that makes me go “hmmmm”….

Rattus Norvegicus says:

Tell the geeks at WordPress that HTMLtidy can fix lots of egregious problems in user entered HTML. There is no excuse.

Wordpress is free, open source software, isn’t it? Then, noone can really blame the geeks at WordPress since users are enabled to fix it themselves.

A solution like TinyMCE is not a good idea. Bad security, heavy on bandwidth and performance. My preference for public facing user input forms is BBCode. There’s a great BBCode parser in PHP called NBBC.
I made a very little, 10 tag javascript RTE plugin for this library and you can convert BBCode to HTML either going into the database(higher performance) or coming out(more extensible). Surely there is something out there like this for WordPress.
Also, a comment preview would help catch alot of this stuff.

Strike one?

I think that I get it… maybe not…

Rattus,
“You are completely wrong. HTML does have valid self closing tags, button, input, hr, br are a few of them.”
HTML doesn’t have self-closing tags, what it has is empty elements that don’t have end tags. They don’t need closing because nothing has been opened. If you tried to close them yourself, it would be invalid HTML.
HTML has elements, some self-contained in a single tag, some with content surrounded by opening and closing tags. XHTML is supposed to be valid XML which tries to handle all tags generically according to the same syntax. All tags come in pairs; an opening tag and a closing tag. The idea of a self-closing tag only makes sense in XHTML where all tags have to be opened/closed.
One thing I did find interesting about this was the fact that because WP publish their code, it’s possible for other people to figure out exactly where and why it goes wrong. The other part of the process – them acknowledging the problem and fixing it – is still missing. But open code does have it’s plus points.

My PHP foo is less than it could be, but it looks like the str_replace function if placed in a new function in your functions.php file would do it for you.
Something along these lines:
add_filter( ‘the_content’, ‘close_tags’, 25 );
function close_tags( $content ) {
return str_replace( ‘i/’, ‘/i’, $content );
}
Any PHP wizards here that could help? I’m sure the above is not 100% correct (namely the content variable I expect needs to be the post variable).

Richard Holle @ 8:06 PM
Your comment suggests an experiment. Test the rankings at each search engine before and after adding those key words. There must be some reason this fairly simple experiment is inadequate, else it would have been done already.
Go, go grantwriters.
======

“Any PHP wizards here that could help?”
I don’t claim to be one, but I thought it might go a bit like the plugin text below. (It’s a million to one shot, but it might just work.) However, I make no claims for the safety or correctness of this. I’ve got no way to test it.
<?php
/*
Plugin Name: Filter malformed tags
Plugin URI:
Description: Replace mistyped i/, b/ and a/ tags in comments
Version: 1.0
Author: Nullius in Verba
Author URI:
License: Lesser GPL
*/
add_filter(‘comment_text’,’filter_malformed_tags’);
function filter_malformed_tags($content) {
$bad_tags = array(‘<i/>’,'<b/>’,'<a/>’);
$good_tags = array(‘</i>’,'</b>’,'</a>’);
$content=str_ireplace($bad_tags,$good_tags,$content);
return $content;
}
?>
Hopefully that won’t mess up the page too badly. Apologies to moderators in advance if it does.

The last time I tried to use the “Test” page (and again just now), there was no “Leave a Reply” section at the bottom of the comments. How do we test comments there?
REPLY: You must not be living right, it is there now 😉
What happens is that WP.com auto closes comments after 30 days, I have to keep reopening – A

The WordPress dashboard has a checkbox “Settings – Writing – WordPress should correct invalidly nested XHTML automatically” that closes unclosed tags in comments, and the setting is available to WordPress.com users who cannot use custom plugins.
WordPress.org users can use a custom plugin to do it more neatly in some special cases, but even then the built-in setting is a useful fallback for those cases that the plugin misses.

Sincere apologies, Anthony. As someone who frequently notices speeling spelling mistakes on newspaper websites – Telegraph take note – I can understand the annoyance.

“Looks like you nailed it Nullius in Verba :)”
Hardly. The plugin shows how trivial it would be to do, but there’s a big difference between WordPress.org, the free host-your-own blogging software, and WordPress.com, the hosting service. Plugins are only allowed on the former, at your own risk. Anthony can’t change any of the code or add plugins, preview tools, or edit controls because WordPress.com won’t allow it. Their buggy software messes up comment handling, and there’s no way he can fix it.
The only solution is for commenters to not mess up their tags.
I’m sure WordPress know perfectly well how to fix it, but apparently they’re not going to. WordPress.com do say that if you want particular features then you can write and request them, and the most popular may be implemented one day if they happen to feel like it, but I wouldn’t hold your breath.

Oh, all right. </i> = “slash italics”. <i/> is slashing something else.

sunsettommy says:
March 20, 2011 at 7:48 am
JoNova,has a very nice comment box set up.
It was set up by a computer expert Steve LeMaster,who was the Founder of Global Warming Skeptics blog/forum,that I now own.
I get the impression that the real problem is that WordPress.com does not allow a fully developed comment box as seen at JoNova.Jo does not use WordPress server.Thus is able to have the nice row of html tags to use that are preset.
Go see what I mean in this link:
http://joannenova.com.au/2011/03/yes-it-is-about-the-science/#comments
Scroll to the bottom of the page and see it.

At the bottom of the JoNova page it says, “Powered by WordPress.”
Maybe she uses a paid version and not the free one that Anthony uses. The Comment box with the select-and-click formatting and a Preview function would be a big improvement here.
/Mr Lynn

WTF! Is this post a Joke?
This is an ‘article’ on the BEST!!!! Climate technology site worthy of comment? Interesting that We the Commenters can ‘break’ the website with a synatax typing error! I would think that some computer program with field validation or range checking rules would have solved this problem. In Teletype days it was called a rubout.
I understand the plea from the moderator. Is this site the best that the Internet technology offers? Hardley! It’s archaic! How many $ will it take to get a state of the art Internet Anthony Watts Website?
Time to move up the ladder! What will it take to move this blog into real time Internet technology?
Yes! I still have a Model 15 Teletype. It works and I can ‘smell’ it, but I have moved on to other technology!
I got a Social Security check for $1000 I’ll donate to a really good Anthony Watts website! Anyone else?
AA4SW

Yeah, preview would be good.
I could live without HTML in my comments, but I surely do like to suppress long, ugly link addresses. And I like blockquotes, the occasional italics, and bold from time to time. So I test it on a Blogger site I have before pasting it here.
That may bite me one day as I’m sure that WP and Blogger aren’t congruent in how they handle HTML, *and* Blogger’s preview isn’t as WYSIWYG as it claims to be.

Eric Ellison says:
March 20, 2011 at 3:28 pm
“Time to move up the ladder! What will it take to move this blog into real time Internet technology?”
Actually WordPress is hosted on a cloud server which is generally considered to be cutting edge stuff. In reality the cutting edginess is economic in nature rather than nifty new usability features. One restriction is that customizing WordPress with plugins on a blog by blog basis is prohibitively difficult. Another restriction is that the blog owner has no access to IP addresses of commenters which is sometimes a very valuable anti-troll tool.

mikemUK says:
March 20, 2011 at 4:06 pm
“The trick I most admire at present is saying what you think, then sticking a thin line erasure through it so it can still be read.”
Yeah, that’s a very gay clever embellishment to be sure!

Ed Dahlgren;
tinyurl.com has a feature called “Custom alias”, so one can put a descriptive title instead of the alphanumeric index. E.g.:
http://tinyurl.com/Zombie-Brains-Memo
😉

Just testing the Italian, or something like that. A bold new adventure.

It’s the Sun dammit, hee, hee. Scepticalscience.com keeps getting it wrong in so many ways. It’s turning into a slow train wreck. Recheck.

You can test it by opening a text file in notepad or equivilent basic text editor. Save the file as “filename.html” and then open it up in a browser.
A proper html file should have a the correct doctype declaration, a head and body etc. but for the purpose of testing simple text formatting, you don’t need to do this.

OK, I guess I don’t know how to do the clever html stuff here, should be simple though, underline, strikethrough, etc.

Two Words: Rich Text Editor
Haven’t I seen WP blogs that use a javascript-based edit box (usually called a “rich text editor”) like the ones provided in most forum input windows? That keeps things rolling along nicely for many a website.
(It’s a Joe Biden joke)

It’s the Sun dammit, hee, hee. Scepticalscience.com keeps getting it wrong in so many ways. It’s turning into a slow train wreck. It really is the Sun and PDO, not CO2 induced CAGW. It’s natural climate variability.

Nope, it does nothing. Superscript O2 and subscript CO2 are there, too.

I just ended up on this site through a link and wanted to inform you that your layout looks a a bit broken, the sidebar looks a little distorted. At least in Netscape, so the problem might be related to my browser.
Just wanted to let you know about it so you could look into it if you want to…

Aura;
Netscape? Netscape?
I got news. It’s not the site which is broken. Your browser comes from the wrong century. Even AOL has given up on it. Load and run Firefox 3.6.4.1, the most recent incarnation of the Netscape legacy.
I’m surprised you can read any sites at all!

K~Bob says:
March 21, 2011 at 5:56 pm

Two Words: Rich Text Editor
Haven’t I seen WP blogs that use a javascript-based edit box (usually called a “rich text editor”) like the ones provided in most forum input windows? That keeps things rolling along nicely for many a website.
(It’s a Joe Biden joke)

Words are always hard to count especially using numbers.
(Yogi Berra might have said something like that)