Slashdot Mirror
Stix Scientific Fonts Reach Beta Release
starseeker writes "At long last, the STIX project has posted a Beta release of their scientific fonts. The mission of the STIX project has been the 'preparation of a comprehensive set of fonts that serve the scientific and engineering community in the process from manuscript creation through final publication, both in electronic and print formats.' The result is a font set containing thousands of characters, and hopefully a font set that will become a staple for scientific publishing. Among other uses, it has long been hoped that this would make the wide scale use of MathML in browsers possible. Despite rather long delays the project has persisted and is now showing concrete results."
chicken
Why exactly was it necessary to link to the user agreement rather than say an example of the fonts or something a tad more useful?
Well good, they needed a font set that had all the symbols you'd ever want to type in science. Only one little problem though...how do you type it? You'd either need a seriously huge keyboard, someone to memorize thousands of key combinations on a current keyboard, or an on screen keyboard program. Each of those options is unacceptably slow or difficult. Plus right now, we have alt codes that almost nobody knows about or uses and the character map built into XP with searchability. So um...what did they invent that we don't already have other than a font?
Google's Super Secret Search Algorithm: SELECT @search_results FROM internet WHERE @search_results = 'good'
Umm, we already have a perfectly good standard solution. It's called TeX.
Can't find anything useful on the website (without giving e-mail address), here's why: <a class="starter" accesskey="5" title="STIX Beta Test" href="#">STIX Beta Test</a>
In university myself and my fellow EE students got through using Maple,Matlab and Equation editor from MS Word. It worked well for us, if someone could have ported the maple into word in a usefull fashion (equation editor is slow at times) that would have been perfect.
The idea of learning a several thousand large charracter set with all the associated keyboard shortcuts holds no interest for me, I'm pretty certain no other engineer on my course would have even attempted it. Perhaps these people could have better spent their time writing a plugin for OpenOffice which gives a highly responsive and good adaptive menu system at the click of a button.
By that I don't mean that the Stix fonts are illegal -- far from it!
What I mean is that the legal profession could use a similarly open-licensed set of fonts for all aspects of the legal process, so that (among other things) it would be one notch* easier to have completely open source case-management / report-creation software at all layers of the legal system. (I'm thinking of American courts, law offices, etc, right now, but not reason why this should apply only within the U.S.)
Something as trivial (and as tangential to content) and which particular font is chosen is one thing that I'd love to see gone. You might be amazed at how difficult it is to computerize even some very busy court systems / law offices (partly because they're busy). I'm doing a clinical at a defender's office with quite a brisk business, and the computer situation is straight out of a Kafka -- lots of PCs are 8 or more years old, there's no reliable Internet service over which to do research (besides which, the computers are so virus-ful that this wouldn't happen anyhow, because browsers don't work on them anyhow. Or, should I say, "browser." Guess which one?) Oh, and installing any superior software is "against policy." Also, offices aroudn the state (New Jersey) are being flopped over to Word, despite everyone preferring an ages-old version of WordPerfect, "for consistency." Goodbye to years of macros -- many documents must be literally retyped.
So most of the above rant has nothing to do with fonts, I realize -- but it does have to do with supporting anything which would ease the replacement of proprietary junkware with something more open on as many fronts as possible.
timothy
* For whatever value of "notch" you think makes sense, that is.
jrnl: http://tinyurl.com/c2l8yr / foes: http://tinyurl.com/ckjno5
I suppose it has something to do with the "openness" of the fonts, or something like that, but haven't complete (or nearly so) scientific font sets been around for a long time? Other posters have mentioned the TeX collections, and there's also Mathematica's fonts: http://support.wolfram.com/mathematica/systems/windows/general/latestfonts.html.
Basically: what's new about the Stix font set?
I'm an amateur typographer and a typophile. I certainly see the need and use for this fontset. However, based on the nature of the comments that I've seen so far, I'm going to sit this discussion out. (Hint: several of you guys are making yourselves look like idiots.)
The one question I have about these fonts is this: Are they designed to sit well in various types of body copy? That is, do the weight and color of the STIX Fonts blend in well with the various serifs and san serifs typefaces used in different scientific publications?
I am glad to see the license for the fonts being published clearly and prominently so it can be reviewed along with the fonts. I recall submitting critique of an earlier license for the fonts, pointing out that the license didn't allow modification (important for improvement) or subsetting (important in PDFs). It was unfortunate that these fonts were aimed at an academic audience, people who were remarkably likely to want to improve the fonts to suit their needs, yet were disallowed from doing so under the old license. The revised license appears to have remedied my issues with their previous license; this license allows modification, subsetting, copying, and distribution (including commercial distribution) all with remarkably mild restrictions that (in my opinion) would not stop this from being a Free Software license.
Because the license allows distribution of the fonts and "the associated documentation files", you could probably find a copy of the font software somewhere that doesn't make you go through a click-through, as well as a sample rendering.
Digital Citizen
It seems like for a lot of the journals out there, it is a Word/Mathtype vs LaTeX world out there. Anyone seen any acceptance of Open Office/Math Editor?
There's nothing new about being able to produce good-looking math output using free software and free fonts; people have been doing that for decades using tex/latex, and the relevant fonts are free enough that they can be distributed with linux distributions.
What's really new and important about STIX is that it will work better with technologies other than latex, especially web browsers. Mathml has been kicking around since 1999, but browser supported has always sucked to high heaven. One of the things holding browsers back from implementing mathml well has been the issue of fonts. Mathml is xml, so it naturally should use unicode. Latex dates back to long before the creation of unicode, so all its fonts are in obscure non-unicode encodings. The approach so far has been to cobble together something that works by building a Frankenstein's monster made out of various fonts that weren't designed to look good together, and that come from various sources. Even though Firefox now has mathml enabled by default, and I have the recommended witches' brew of fonts installed on my linux box, firefox still nags me about its fonts every time it needs to render mathml. The only way this is going to get better is with the STIX fonts.
For an example of how screwed up things have been, take a look at the archives of the Wikiproject Mathematics talk page on Wikipedia. WP's software uses software that renders LaTeX math into bitmaps, and that software has only very limited mathml output functionality, which is not actually being used. There was a project by a math grad student at harvard to make something better, called blahtex, which would have allowed mathml to be output as well. A user who was interested in mathematical topics, and who had Firefox, could set a preference on his WP account so that math would always be displayed to him in mathml, which would look much better (both on the screen and on paper) than the crappy screen-resolution bitmaps. Well, he wrote the thing, got it working great, tested it extensively on a huge number of equations harvested from actual WP pages, built support for it among WP editors. And when all was said and done, the Mediawiki developers wouldn't take his code. Basically the reasoning seems to have been that browser support for mathml sucked, so there was no point in disturbing mediawiki's codebase for a feature nobody cared about.
Ouch.
It's been a real chicken-and-egg thing. Since mathml support in IE requires a plugin, nobody's bothered to put much effort into making mathml content. MS's motivation for building mathml support into IE has been low, because nobody was using mathml, and the fonts weren't available. Although firefox has mathml support, it's extremely buggy, and the motivation to fix the bugs has been low, because nobody was using mathml, and the fonts weren't available. The fact that STIX is finally coming out may finally generate some excitement among developers about making mathml into a going concern on the web.
Anothing thing holding everyone back is that people are still expecting to be able to write html as if it was 1995, with no quotes around attributes, unbalanced tags, etc. That isn't going to work for xml-based technologies like mathml, and in fact firefox won't render mathml if it occurs on a page that's not valid xhtml. That seems to have been one of the big factors holding back adoption of mathml by mediawiki, for example, because the html code generated by mediawiki isn't valid xml.
I'm really hoping that sometime soon square roots won't look messed up on the screen in firefox's rendering of mathml, and a printed mathml web page won't look so horrible.
Find free books.
They will have a big job replacing the computer modern fonts, especially if they don't make convenient LaTeX packeges to load the fonts.
SJW n. One who posts facts.
Compare these to the fonts of yore, such as Times or Arial or essentially any font that existed in the early Mac and Windows days. The font designers took great care to ensure that bit maps were customized for best appearance at small point sizes, given the inherent limitation of the black-and-white screens and resolution available then.
Now it seems it is universally assumed that everyone will have smoothing turned on. Modern fonts may look professional and polished at larger point sizes, but the unsmoothed bitmap versions of many of them at small sizes tend to look rough and amateurish, with ugly artifacts and inconsistent line widths and sometimes barely legible. Even the smoothed ones aren't necessarily great at small sizes - the smoothing can make them blurry with poor contrast, unlike the crisp black and white of well-designed bitmaps.
Perhaps I am alone, but I am more efficient working with small font sizes for things like programming, so I can have the maximum amount of information simultaneously available on the screen. So I almost always have smoothing turned off and use old-fashioned (and typically mono) fonts that have clean, carefully crafted bitmaps suited for that purpose. But when I switch to web browsing, if the site sports a trendy font and I have smoothing turned off, it can be an eyesore.
Can anyone point at a good reference.
I'm familiar with Type1, Postscript, bitmap, TrueType; but not OTF.
-- Rich
Some of the fonts apparently crash FontBook when previewed. It's too bad, since I was hoping for a good symbol font.
I downloaded the STIX zipball and glanced through some of the character sets included therein. The fonts are very attractive and I think that within a short time of delivery, many technical publications will adopt them. It is only a shame that this project has suffered so many long delays. It's almost like waiting for Sarge to get released.
If anything can do it, it'll be an initiative something like the STIX work.
In any case, Computer Modern is far from everyone's taste. Knuth did a great job designing a highly legible font that could both typeset mathematics elegantly and survive the scanning, photocopying and other abuse scientific papers tend to suffer. However, notwithstanding Knuth's personal preferences, aesthetically the Computer Modern set leaves a lot to be desired. Many people prefer a different style on paper, and on screen the lightness of the CM set is pretty horrible, as anyone who's tried to read a long PDF of a paper set using TeX can testify.
It's a shame that in a world where OpenType and Unicode are now commonplace, and where many professional fonts now come with glyphs for numerous different alphabets and numerous carefully tuned typographical features, it isn't yet common to supply matching glyphs for say the top 100 scientific symbols. I guess the market is just too specialised, and the current dominance of the TeX family means there's little commercial incentive for others to produce high quality scientific fonts. In that respect, having a high-quality, science-friendly font available for use with things like web pages surely must be a good thing. (Monospace fonts useful for typesetting computer programs currently suffer a similar lack of support, probably for the same reason.)
If you disagree, post your argument. (-1, Overrated) isn't your personal censorship tool for views you don't like.
The project history is sort of amusing.. Originally scheduled for release in Summer 2005, the release date was delayed to Sept 2005, the Dec, then March 2006, then June. In July 2006 it was announced the fonts would be ready "in two weeks". Every two weeks since then, they've made announcements that it would be ready in two weeks more. (literally.. see their news page). Anyway, they must be happy it's almost done!
Now we can get our hopes up for DNF - the Stix fonts have arrived!
Oh, and some place named Hell called wanting to place an order for some heating units...
I love how none of the links on their web site's menu work.
Much confidence is inspired by a website that does not work with Firefox, and wants to harvest your email address to allow you downloading their beta software (well, font).
I really don't understand what all the hate is about for Computer Modern. I think it's a fine font.
Of course, it doesn't look nice on the screen when viewed at 100%. But that's what you get for viewing something at 72 (or 96 or whatever Windows uses) dpi, that's designed to be viewed on paper at 300dpi.
If you blow Computer Modern up to 150% or so, which in my experience tends to be what happens if you fit the width of a document to a good-sized monitor, I think it looks pretty good. But at 10 or 12pt at 100% magnification on a low-resolution device like a computer monitor, you lose all the fine detail that you need.
I guess I can see how there's a need for an alternative for people who are doing all-digital workflows, but if you're going to print the paper out at the end, there's nothing wrong with Computer Modern. It certainly beats the pants off of Times New Roman, IMHO, if you have a good 300+ dpi laser to use as an output device.
"Ladies and gentlemen, my killbot features Lotus Notes and a machine gun. It is the finest available."
Already in Fedora, on the assumption 4. is 4.a or 4.b (making it a Free/Libre font according to us)
https://fedoraproject.org/wiki/SIGs/Fonts/Triaging/Pipeline#head-a970f733a2659c3045c01321c4f775536fa0ff8f
To me it seems like both projects took ridiculously long to complete and they were delayed over and over and over again. Anyway, for better or for worse now they are both here ... in beta ;)
If you don't fail at least 90 percent of the time, you're not aiming high enough. (Alan Kay)
I've done a fair amount of statistical publishing, and have found that most statistics journals (at least the ones I've submitted to) will accept pdf or ps files. I feel like those are usually better anyway, because there's no question about what was intended to appear (usually).
I actually use Openoffice now for everything (not just statistics papers), and not Word, because of the equation editing issue.
I've tried Tex, and it's great, but too difficult to use for things _not_ involving heavy equations.
Openoffice is a nice compromise, in my opinion.
For many working scientists, manuscript creation is performed using a monospaced typeface.
I admit that I was initially quite excited when I downloaded these files a few days ago. (Yeah, I lead a sheltered life: a new typeface can excite me.) The excitement evaporated when I realised that there seems to be no monospaced typeface. I might consider using these files for final output -- I'll have to see how it looks in practice; the individual glyphs look quite nice -- but this announcement hardly supports the entire process as their mission statement suggests.
Try Lyx, a GUI for LaTex.
Word 2007 has a completely revamped equation editor and includes a new font specifically designed for laying out math equations (Cambria Math).
If you haven't given it a whirl, you should. Quick, easy, approachable, and it produces beautiful output. You can even cut/paste equations to/from MathML.
Neil
Tell me about scientists without an agenda of their own. Religion was and is always used as a means of improving things. Why do you think people know how to read and write? It's only because someone thought it helpful to write down religious scriptures. There's no chance of dividing religion from science. Whatever SIL does, they certainly improved linguistics as well (even if I only happen to use their superior Gentium fonts-which happen to be more suitable for linguistics than Stix). Speaking of linguisics and religion, tell me what Arabic would be w/o religion. Islam created Arabic (at least as we know it). The language people think to speak (colloquial Arabic often called) has not much in common with what Arabic is considered by the Arabs themselves. Highly interesting, religion and language.
Make OpenOffice accept these fonts first. OOo does not like .otf fonts, at least not on Linux. .pdf, see e.g.
And it does not export them to
http://www.openoffice.org/issues/show_bug.cgi?id=43029
http://qa.openoffice.org/issues/show_bug.cgi?id=78858
I hate fonts that allow O and 0, and 1 and l to look so similar. This has only been a major headache in the computing world for a few decades, no need to change now. I don't care if they are experts, it is stupid and short sighted to continue this mistake. Before you say anything, I do use a font that sensibly corrects this error.
Is he strong? Listen bud, He's got radioactive blood.