ArticlesBack in the old days {When? Describe what technology was like in those days to give people a feel}, I transcribing lots of articles about γ-Aminobutyric acid (GABA) {reference?}, so I memorised the ALT-keycode{what’s the proper name?} combination {What’s the magic number and describe the process} for γ {will this lower-case gamma work on other people’s browsers? Wait, that’s upper case! Go check, and reference the Unicode #.}, but these days I have to search Google {add a link to the search term} for a character so I can cut-and-paste {shouldn’t it be copy-and-paste?} it.
These days, I use non-English characters rarely – perhaps once in every 50,000 words I type? {Find a way to check that stat.} The problem is that it is so hard to include such characters with an English keyboard in Windows. If it was easy, my use would sky-rocket to once every 5,000 words I type. {Bold claim. Doubt it is true. How will you support it? Try using Sapir-Whorf to baffle brains; you could show that having more foreign characters available makes foreign concepts easier to pick up. Beautiful, but is that how you spell Sapir-Whorf?}
Charmap {I bet that’s not the right name. I bet is it Character Map Tool, or something like that} is an application shipped with Windows {which versions? is it in Windows 7?} which allows you page page down through the thousands {estimate} of characters {Is character the right word? I bet it is called a pictogram or something like that} available in your font {or is it font-face? Whatever.}
In order to insert a character, you have to go through a long sequence of keystrokes {I don’t know if there a shortcut?} including paging through tiny pictogram after tiny pictogram {?}. When you select it, it gets put in a buffer. Then you press copy and it is in your clipboard to paste into another application.
If we have learn one thing from Google and Windows Vista {Oooh, I bet the Apple-heads will point out this was a feature of the original Macintosh. Better find out when it appeared in their world.} it is that searching is the best interface {best? prove it, or reduce the claim} way to find a single item from hundreds. {Is that really the best distinction for when search is best? Citation needed. Soften the claim by mentioning that you need to have aliases recorded too, or people will point out that if you don’t know the right word you’ll never find it.}
The Pictogram Map Tool Utility Application (PMTUA) {Or whatever Charmap is officially called in Windows. Add a link too, for people who want to learn more!} of the Future should be quite different.
{Oh, and research IME – whatever that stands for, and make sure it doesn’t already do what you are asking for, or you will look like an ignorant monolinguistic fool.}
Unlike today {? Double-check, especially Windows 7?}, PMTUA Of The Future should have a hot key to launch quickly – let me emphasize that! – quickly, and offer a text box to type the name of the character – “degrees” will get you °; “pi” will get you Π and π. “Arrows” will get you a large selection of options.
When you find the right charactergram, you have several options:
- select to get you a larger version, it’s official name, perhaps a way of finding out whether it is available in a given font, and perhaps some related characters – so finding the white happy face, will show you where to find the black happy face and the white frowny face.
- select and copy (including CTRL-C) will copy that character into the clipboard, and then get the hell out of your face. The task is done. You are finished with the app. Have it close (perhaps remembering where your context were for the next run).
- double click will push that character out as typed input to the last window you were editing, and, once again, get the hell out of your face. Easy!
As you can see, this well-researched article proves beyond doubt that my key ideas (searchable characters, and faster selection) are the most important ones for Microsoft to work on next, and if they would only listen to me, the world would get better and better, even if it is only one ° at a time.
Comment by Aristotle Pagaltzis on June 2, 2010
Sorry Julian, but this one did not work out. The blizzard of parentheticals is so dense it obliterates any chance of ever making it to the end of a sentence while still remembering what it said. I had to delete them all after copy-pasting into a text editor before I could actually read the post. (Let it not be said that you don’t have dedicated readers.)
The idea I found after slashing my way through the textual undergrowth is nice. I think I saw several websites, actually, that implemented quite similar interfaces using Javascript. There was a cute one where you search by visual example, by typing some (even vaguely) similar-looking character (eg. typing “c†will yield such graphemes as “⊄â€). {In keeping with the spirit of your entry, I did not bother to look up any links.} Of course websites are an unsatisfying approximation of a good desktop application for a purpose like this… but that they exist shows that the idea is sound.
Comment by Julian on June 2, 2010
Ah well. Experiment attempted and negative result. Thanks for the feedback (and your dedication!) I briefly considered coming up with a CSS style to make words in braces easier to gloss over, but spending that effort didn’t quite seem consistent with the message.
I’ll try to find some of these web-sites. Maybe I can add my own hotkey to launch them.
Comment by Julian on June 2, 2010
Here is the best online Unicode character map I found.
It allows some searching, but it isn’t great. Try searching for “pi”, and you don’t get Π listed, just π. Try again with “pi “ and it just gets weird. You need to search for “Pi “ (note case; you know how I feel about case-sensitivity). Searching for “gamma” works better. “Degree” works well but “Degrees” misses the symbol I was after.
Comment by Aristotle Pagaltzis on June 2, 2010
In keeping with the spirit of your last comment: ;-)
I looked for and found the visual character lookup page I mentioned.
Comment by John Y. on June 2, 2010
I was able to read this article as-is, though rather slowly and with multiple visual passes per sentence. I don’t know whether this means I am more dedicated or less dedicated than Aristotle. 😉
Taking a very different and ambitious approach, there is at least one Web site which offers handwriting recognition, where you draw a symbol and it finds the proper LaTeX for what it thinks you drew. While this isn’t what you’re describing, I can see how it could be fleshed out, mapped to Unicode, and developed into a fairly rich PMTUA. And, since it looks like we’re including links now, here is the one I’m aware of.
Comment by configurator on June 2, 2010
Erm, charmap has search. “Degree” found “°”. “Pi”, unfortunately, is a substring of “Capital” so it found all capital letters. But “Omega” found “ÎÎ©Ï‰ÏŽÑ Ñ¡ÑºÑ»Ñ¼Ñ½á½ á½¡á½¢á½£á½¤á½¥á½¦á½§á½¨á½©á½ªá½«á½¬á½á½®á½¯á½¼ÏŽá¾ ᾡᾢᾣᾤᾥᾦᾧᾨᾩᾪᾫᾬá¾á¾®á¾¯á¿²á¿³á¿´á¿¶á¿·á¿ºÎῼɷ”, which is pretty comprehensive.
Comment by configurator on June 2, 2010
Oh, and I forgot to mention it gives you the unicode value and official name as well: ῼ is “U+1FFC: Greek Capital Letter Omega With Prosgegrammeni”
Comment by Julian on June 2, 2010
I checked out the link Aristotle provided. It finds HTML entities only, which may be why when I type “c”, it doesn’t find ĉ. Nor did “snowman” find any of the snowmen characters. I thought the look-alike feature would allow me to find ∞ by searching for “8”, but no luck.
John Y’s drawing site is pretty cool. It is limited to LaTeX characters. I tested it on the Tau, Epsilon, Chi from LaΤΕΧ and it (only) took a couple of tries. On the other hand, if you can get it to return a snowman, you are a better drawer than me. (Not even sure they appear in the list.)
Comment by Julian on June 2, 2010
@configurator,
Charmap has search? Ha! That is hilarious! I am *so* glad I highlighted my lack of research, or that would have been very embarrassing!
Sure enough there is a little “Advanced View” checkbox that I have never noticed before. {Check when it was added, maybe it only appeared recently?}
In keeping with the rest of the application, the usability is awful!
No incremental search.
Try searching for “smilie”. No responses. Now change it to “face”. Whoops, I got everything returned. And my search term disappeared. What happened?
Well, when you search, the Search button turns into a Reset button, and you have to press that before you can search again. But the search term is not greyed-out, so you can edit it (with no effect) but when you hit Enter, it resets it, throwing away what you typed.
Now, search for “face” again. Why only two faces returned? Unicode has at least three smilies plus 6 dice faces, and some others. Ah, it is because the current font (Arial) only has two. Arial Unicode MS, by comparison, has five matching characters. Terminal doesn’t let you search at all.
Thumbs down for Character Map (as it is officially called.)
Comment by Aristotle Pagaltzis on June 2, 2010
Yeah, I think the lookalikes lookup can be improved. Note that it also answers some queries by meaning rather than looks, ie. you get “π†and “Π†if you ask for “piâ€. The data set for both kinds of query could be expanded a lot further, that site doesn’t seem even close to exhausting the possibilities of the concept.
Comment by configurator on June 3, 2010
I never said usability is good, it’s utterly rubbish. But it does do all those things you wanted 🙂
I think part of what charmap is meant for is only showing you the characters available in a specific font. Why don’t you develop your own charmap, then?
Comment by Julian on June 3, 2010
Once you get past the ironic poor usability of my writing style, the article did talk about the desired minimalism of the PMTUA of the Future, to improve efficiency of using the app. So, it doesn’t quite do everything I wanted.
Why don’t I develop my own? Partly because my demand isn’t big enough to warrant my effort. Partly because I am already stepping up to the implicit challenge set down by Vinay, the maintainer of the Python logging component, to provide a re-usable component to solve that problem. Look for an announcement soon.
Comment by Sunny Kalsi on June 3, 2010
When I did my talk on OpenGL, there was a lot of maths, and there’s a Javascript library which lets you type like in LaTeX, but then converts it into the appropriate characters (http://www1.chapman.edu/~jipsen/mathml/asciimath.html).
Linux has an input API. It used to be called SCIM, but the new one is called IBus. When I used the SCIM mode for Kanji, I could type “katakana in japanese is [Ctrl-Space]katakana[space to choose alternatives][enter] but [Ctrl-Space]hiragana[space to choose alternatives][enter] is more commonly used” and it would do what you expect. The new methods I think do it in stride so you don’t need the extraneous control characters. Here’s something I wrote when I was playing with this stuff: http://blog.quaddmg.com/2004/06/25/nuclear-proliferation-aka-mr-glorious-is-not-feeling-well-%E3%81%8B%E3%81%8F%E3%81%8B%E3%81%8F%E3%81%95%E3%82%93-%E3%81%92%E3%82%93%E3%81%8D%E3%81%AA%E3%81%84
I don’t know of a “maths” language, but I imagine it would work exactly as you describe it.
Comment by DeeJuggle on June 5, 2010
Thanks for inspiring me to have a play with the Windows (XP) Japanese IME (‘Input Method Editor’). It produced Greek letters (both upper & lower case), mathematical symbols and simple shapes just by typing the name of the character/shape directly into the document.
Comment by Julie Lawrence on June 28, 2010
http://copypastecharacter.com/ – small set of characters, but nice and simple
Comment by Julie Lawrence on June 28, 2010
Re “I briefly considered coming up with a CSS style to make words in braces easier to gloss over”, check out http://www.telescopictext.com/