Talk:ISO 31-11

From Wikipedia, the free encyclopedia

Contents 1 Unicode UP TACK versus PERPENDICULAR 2 Angle brackets 3 Note on Unicode repertoire used 4 Specificity of ISO 31-11?

[edit] Unicode UP TACK versus PERPENDICULAR

Okay, now we are getting really deep into the Unicode standard, but shouldn't the perpendicular operator be ⟂ (U+27C2 PERPENDICULAR) instead of ⏊ (U+23CA DENTISTRY SYMBOL LIGHT UP AND HORIZONTAL, as currently used in the article)? --Abdull 22:15, 28 May 2006 (UTC)

Well spotted! My mistake. The character I actually wanted to use was ⊥ (U+22A5 UP TACK). I took a wrong turn in the code charts, was heading for APL symbols, but actually ended up in Japanese dendistry symbols ... a likely peril in the age of mega character sets. UP TACK is widely supported in existing fonts and seems the appropriate Unicode 4.0 character for the perpendicular symbol to me. A dedicated ⟂ (U+27C2 PERPENDICULAR) was added only quite recently (Unicode 4.1) and I'm not fully convinced yet that there was a good case to disambiguate between the two. The glyphs for UP TACK and PERPENDICULAR are for all practical purposes identical, and the difference between a constant and an operator only matters in TeX-style math-rendering implementations that try to automatically insert whitespace near operators, which is irrelevant for HTML, Wikipedia, etc. So I prefer to use UP TACK instead of PERPENDICULAR for the moment, because the latter is not yet widely implemented in fonts. Markus Kuhn 19:19, 29 May 2006 (UTC)

[edit] Angle brackets

I'm seeing ? where I think other character should be. In the mathematical symbols table it has the line:

⟨⟩ |?a+b?c | ac+bc, angle brackets

Shouldn't that read:

<> |<a+b>c | ac+bc, angle brackets

Or am I missing something.

The font your particular browser uses probably still lacks glyphs for the two characters U+27E8 (MATHEMATICAL LEFT ANGLE BRACKET) and U+27E9 (MATHEMATICAL RIGHT ANGLE BRACKET). The angle brackets have, unfortunately, a somewhat confused history in Unicode. They were for a while mistaken to be CJK wide characters, and disambiguated from these (unnecessarily) rather late. Markus Kuhn 22:01, 3 July 2006 (UTC)

It does seem to be a font issue, but I wonder if many people will have the same issue. I'm on Mac OS X 10.4.6 and though I have all of the standard internet/web fonts installed (and then some) I still don't have any that have a glyph for this character. I wonder if a note should be added to show what it should roughly look like using another more commonly existing glyph (like simply the greater than less than charaacter glyphs like I used above). I don't think the characters in the main text should be changed, but just suggest a note be added, since I don't imagine I'm the only one to encounter this problem. As the fonts improve and proliferate, the note could then be taken away. --Cplot 23:13, 3 July 2006 (UTC)

BTW, I have the same problem with the perpindicular glyph in the comment above (though the dentistry glyph displays fine, so maybe a similar not should be added there if the peripindicular character is correct).

Please report what exact browser and font you are using, and I'll be happy to investigate and report the problem upstream. I'd rather see this fixed in the font and not in the article. Markus Kuhn 08:41, 5 July 2006 (UTC)

[edit] Note on Unicode repertoire used

Please do not deface the article with lots of commentary on the font inadequacies of individual web browsers and fall-back notations. If at least one major browser (e.g., Firefox under a recent Linux distribution) displays the Unicode character adequately, then this should be sufficient here. I very much like this article to be a bit leading-edge with regard to how Unicode can be used on Wikipedia to help with mathematical and scientific typesetting. Use this article as a reference in bug reports to your font or browser developers, as a benchmark to drive forward the state of the art, and as a copy-and-paste source for authoring other articles. Do not dumb its repertoire down to the oldest web browser that you can find. I'd like to see the Unicode repertoire needed to typeset ISO 31-11 to become commonly supported in fonts. This was my main motivation for contributing this article. The repertoire is already widely supported. For example, all the characters currently in the article have been available for many years in the standard X11 fonts, and are displayed correctly even in ancient command-line text-mode browsers. (If circumstances outside your control force you use outdated commercial products, rest assured that they will catch up three versions later, as usual.) Markus Kuhn 08:41, 5 July 2006 (UTC)

I'm sorry you felt those notes were defacement. I felt I was improving the article, which btw is quite helpful. I fully support your goals for this article, but feel some explicit statements to that effect would be helpful. I'm running another leading edge operating system with X11 installed and I'm missing glyphs for these characters. It took me some time to figure out what was going on and I felt that, rather than make everyone else go through that same process, I'd provide the answers I gleaned for everyone else in notes to the table.

I've sinced moved those note to footnotes so they won't clutter the table. Let me explain my reasons. If we want this to be leading edge, then we need to put the correct unicode character in the text. However, since it's also an enyclopedia article, the general reader should learn the basics about the ISO31-11 standard. If an important glyph is missing then they'll go away scratching their heads. Apparantly neither of us have a perpindicular glyph, so it's counterproductive in trying to push the edge to simply substitute a similar character without noting that fact. It makes everyone reading the article complacent with their font support even when they're missing support for an important unicode character. The characters are important in many more places than just TeX. They're meant to convey meaning: both for machine processing and human readable form in a variety of places. Substituting a character with a similar looking glyph may convey the meaning in human readable form, but it destorys the meaning in terms of machine processing. Finally, on the angle brackets, I think some substitute is necessary. I've checked across Firefox and Safari (a KHTML descendant) and neither displays these characters correctly. As I mentioned I have a standard X11 enviornment installed, so that doesn't help the situation. I also have the standard "web" fonts installed, which is all we should count on a site like Wikipedia. It is probably these fonts that need support added for the perpindicular and the angle brackets.

So to further our worthy goals, I've added these notes again as footnotes. I've also added a requesti for readers to contact their vendors to imrpove support for these characters. I think this furthers our goals.

One last point. i think it might be helpful to list the character references for these unicode characters. In the notes I added I used the character references in my markup. I think that's helpful for someone who's used to looking at HTML source. A step further might be to add a column to the tables with the character reference listed explicitly.

Just a few suggestion. Any thoughts? --Cplot 17:41, 5 July 2006 (UTC)

A couple more comments.To the note, I added the angle brackets from the CJK punctuation category. These look very similar in glyph form and from the name I gather are meant to dennote the nearly same thing in character form. Also I wonder if the note I wrote and the one on unicode support already in the article should not be revised and moved to the top to warn readers of these unavoidable issues? --Cplot 17:50, 5 July 2006 (UTC)

[edit] Specificity of ISO 31-11?

I was wondering too, whether ISO 31-11 actually specifies the Unicode characters or some other character specification? Or does it simply designate a standard of glyphs for use in math notation? It might help the article to make that clear too. I'm sorry if this question is already answered and I missed it.--Cplot 20:29, 5 July 2006 (UTC)

ISO 31-11:1992 makes no reference to Unicode, ISO 10646, or any other particular coded character-set standard. It simply presents example glyphs on paper, very similar to what you get if you printed out this article on a printer using a carefully selected font. Therefore, it would be inappropriate, in my optinion, to clutter this article with a lot of Unicode references, because this is currently beyond the scope of ISO 31. Such a cross-reference might, however, be a good extension for the next edition of that standard. Until then, I think all we should do is make sure that we use the Unicode repertoire that we have wisely. Detailed guidelines on the use of the Unicode repertoire for mathematical typesetting in HTML and Wikis should better be collected elsewhere. Markus Kuhn 13:29, 9 July 2006 (UTC)

That's precisely the type of information that I think would be useful for the article: that the standard specifies glyphs and not characters. It's important to explain to readers that ISO 31-11 recommends representations, and that those representations do not mean the same thing as the unicode characters used in the article. The ISO writes their specifications largely for print distribution. This issue doesn't arise for them. However, here where we're writing about the standard largely for electronic distribution, we face dealing with this ambiguity. It may be the ISO should have dealt wit the ambiguity for the electronic distributions of the proposal, but that's another issue.

If the standard only specifies representations, then we might as well use the greater-than and less than signs for (angle brackets or the CJK angle brackets) or whatever characters will reach a wide wikipedia audience. Those don't dumb down the article, they just serve as approximate representations to what the standard recommends (which is all anyone can do with the level of specificity of the standard).

So to sum up, I think some mention that the standard specifies representations and not precise characters should be made. Then as far as representing those characters we need to follow Wikipedia manual of style and ensure this article is readable by a broad range of readers (i.,e., uses characters for the least common denominator). This gets in the way of the other goals mentioned above. I think it would be far more desirable to typeset the with the appropriate and precise unicode characters for others to use and have a note specifying dumbed down characters than to have the only the dumbed down characters in the article. That's my feeling on that anyway. --Cplot 19:43, 9 July 2006 (UTC)