The Icon Bar: News and features: Getting Unicode Working With RISC OS 4 (and 6?)
Posted by Jon Robinson on 23:30, 21/7/2011
| RISC OS, RISCOS Ltd, Tutorials
What is Unicode? Unicode has been developed as a standard way of representing the characters in all the world’s basic languages. It was designed to make it easier to exchange files, between computer users who write using different alphabets, and to make documents that do not use the Western, Latin alphabet, readable all over the world, regardless of which program, or operating system is used to view them with. Before Unicode was introduced, people used standard eight bit (one byte) fonts, which allowed the representation of only 256 different characters. When creating complex documents which included, for example, special symbols and foreign-language characters, you would have to use several different fonts to get all the required characters. The problems would start when you emailed this to somebody else, and they didn’t have the same fonts installed on their system, that you had used to create it with. Parts of the document would be unreadable. A further problem arises when two different encodings are used for the same set of characters. For instance, Cyrillic web pages are often encoded in either KO18-R, or Windows 1251, which both have the same characters, but in different positions in the font table. If you sent a KO18-R document, or email, to somebody whose system was set up for Win 1251, they might not be able to read it. Unicode was designed to get around these problems by using more than eight bits to represent a character. By using more than one byte, virtually every symbol, or character that exists in any language, can be allocated its own, unique number (or code point), so that the character can be represented by the same number, whatever operating system, or program you are using anywhere in the world. If your system supports Unicode, a word processor or web browser that loads a Unicode document, will check all the code points in the document against a look up table, which tells it whether that character is defined in any of its installed fonts. If so, it retrieves the character’s details, and renders it to the screen. Unicode makes the sharing of documents and web pages, across national boundaries, much easier, and is a God-send to those who regularly communicate with people who don’t use our alphabet, or are interested in studying foreign languages. So how does RISC OS bear up to the challenge of supporting this international standard ? To Install Unicode on RISC OS 4- First download a copy of the Cyberbit font from here.
- Now download a copy of John-Mark Bell’s font converter from here. Extract the program using !SparkPlug.
- Load TTF2f, and Select-click on its icon to open the dialogue box.
- Drag the Cyberbit file icon into the top writeable box, and change the RISC OS name to Cyberbit.
- Click on the drop-down menu, and select !Boot.Resources!Fonts as the destination for the converted font.
- Now click on the Convert button.
BE WARNED, however, it takes about ten minutes to complete on a StrongARM RiscPC, and, for the first two or three minutes, nothing appears to be happening! Do NOT assume that the program has crashed, and hit the On/Off button. Just be patient, and wait for it to finish. - Now download the Fonts (Unicode Font Manager) module, from the Video section of the downloads area of the RISC OS Open website, and get ROMFonts as well.
- Extract Fonts and RomFonts, using SparkPlug.
- Find the !Boot application on your hard drive, hold the Shift key down, and double click on it.
Double click on Choices, Users, Single, Boot and then PreDesk. Drag Fonts and ROMFonts into PreDesk. - Create a little Obey file with the following two lines -
RMEnsure ROMFonts 0.75 RMLoad <Obey$Dir>.ROMFonts RMEnsure FontManager 3.66 RMLoad <Obey$Dir>.Fonts and save it in PreDesk. - Do Ctrl-Shift-F12, (or select Shutdown from the Task icon on the iconbar), to reboot your computer.
- Now, load NetSurf and head on over to the webpage of Nippon Television, to check you’ve got it working properly.
If so, you should see something like this:
Unicode and !PDFIn quite an important development, Chris Gransden has released a new version of his popular PDF reader (version 3.02), which can display Unicode characters in PDF files. To do this, you have to click on the Render as Sprite icon on the PDF toolbar, and wait for the current page to open in Paint. Previously, you would just have seen rows of ‘garbage’ where the Unicode characters were. There is still the problem that the resulting sprite file is not nearly as useful for high-quality printing, as a draw file would be. But it is an important step forward. Getting Unicode Text into a Word ProcessorAlthough at the moment, NetSurf is the only major program on RISC OS, that fully supports Unicode, there is one very useful trick that allows you to get Unicode text off the web, and into a word processor, like EasiWriter. With the required page loaded in NetSurf, click Menu, and then go Page->Export->Draw. Load the Draw version of the page, and click on it with the Selection tool. At the moment, everything on the page is grouped into one big object. But, if you hold down the Control key and press U a few times, the page will begin to break up into separate objects, which allows you to select only the part of the screen that you want, with the Selection tool. Now click Menu, and do File->Save->Selection, to save just the part of the screen you want to use. If you want to change the size of the text, double click on the new file, select the text and use Transform->Magnify to do so. Et viola! The resulting Draw file can now be dropped into EasiWriter. Unicode imported into EasiWriter, courtesy of Draw At the moment, we still cannot edit the text, but by using NetSurf’s Export Draw capabilities, we can effectively cut and paste from the internet into EasiWriter. And because Draw is a vector format, we can not only resize the Unicode text, but should also get much better print quality than we would, if we simply snap-shotted part of the screen with Paint, and used the resulting Sprite file in EasiWriter. For The FutureAlthough viewing Unicode documents is becoming much easier for RISC OS users, thanks to the efforts of Chris Gransden and the NetSurf team, we still have a long way to go. We still have no easy way of actually creating Unicode content. What we really need is a !Chars type application, which can be configured to present useful sets of all the Unicode characters, and which allows users to enter them into a document by clicking on a palette. (Luckily, since I started writing this article, Paul Sprangers in the Netherlands, has started work on such a program, called KeyMap). What we also need, is for at least one of the RISC OS word processors (TechWriter being the obvious candidate) to be upgraded, so that it can handle Unicode characters internally, and save Unicode documents as well. At the moment, we can extract blocks of Unicode text out of NetSurf, and drop them into TechWriter, but we cannot edit them. However, if enough users request this as a feature they’d like to see, maybe this will change . . . Finally, don’t take continued software development on RISC OS for granted. If you regularly download the latest version of NetSurf from their web site, and bits and pieces from RISC OS Open, make sure you make a donation to one of the bounties set up by the latter, to finance further development of the operating system. And remember to buy one of the £5 CDs from NetSurf, next time you see them at a RISC OS show! If you want software development on RISC OS to continue, find a way of showing your support for it. Many thanks to John-Mark Bell of the NetSurf team, and Steve Revill of RISC OS Open for help researching this article.
|
Getting Unicode Working With RISC OS 4 (and 6?) |
|
wpb (20:04 22/7/2011) wuerthne (20:45 22/7/2011) VincentVega (18:55 23/7/2011) pnyoung (11:07 24/7/2011) pnyoung (11:09 24/7/2011) arawnsley (12:52 24/7/2011) pnyoung (09:29 27/7/2011) apacketofsweets (10:50 27/7/2011) trevj (12:46 27/7/2011) pnyoung (14:50 3/8/2011) Reboot (12:39 9/9/2011) nunfetishist (12:47 9/9/2011) apacketofsweets (12:47 9/9/2011) nunfetishist (13:22 9/9/2011) Reboot (12:47 13/9/2011) nunfetishist (13:27 13/9/2011) Reboot (21:34 14/9/2011) pwx (21:55 14/9/2011) apacketofsweets (23:46 14/9/2011) Reboot (15:53 2/10/2011)
|
|
W. P. B. |
Message #118358, posted by wpb at 20:04, 22/7/2011 |
Member
Posts: 22
|
Great article - thanks! I'd be interested to know why PDF can only render its unicode fonts as a sprite. Doesn't it just use the Font Manager for rendering its fonts?
Sargasso also uses RUfl, I believe, though I've never tested to see if I can view anything obscure!
Parmesan is also Unicode-aware, I think. But it doesn't use RUfl, so its rendering of glyphs might be a bit lacking at times.
We've also been discussing the input side of things over on the RISC OS Open site. See this thread about !Chars: https://www.riscosopen.org/forum/forums/3/topics/700 |
|
[ Log in to reply ] |
|
Martin Wuerthner |
Message #118361, posted by wuerthne at 20:45, 22/7/2011, in reply to message #118358 |
Member
Posts: 1
|
That entire section about !PDF looks like nonsense to me. Installing the Unicode FontManager does not make any difference to what !PDF displays: When rendering to a Draw file it uses the FontManager, but not its Unicode facilities. When rendering to a Sprite, it does not even use the RISC OS FontManager at all. That Arabic document can be rendered just fine by !PDF in Sprite mode even without having the Unicode FontManager installed. |
|
[ Log in to reply ] |
|
VV |
Message #118367, posted by VincentVega at 18:55, 23/7/2011, in reply to message #118361 |
Member
Posts: 7
|
Tactful as ever, I see. |
|
[ Log in to reply ] |
|
Peter Young |
Message #118368, posted by pnyoung at 11:07, 24/7/2011, in reply to message #118358 |
Member
Posts: 6
|
Tried it with RISC OS 4 (admittedly on VRPC), where PreDesk is in a slightly different place, and in which the Obey file needs a bit of tinkering. After that, I got a "Font manager in use" message. Probably doing something wrong!
Peter. |
|
[ Log in to reply ] |
|
Peter Young |
Message #118369, posted by pnyoung at 11:09, 24/7/2011, in reply to message #118368 |
Member
Posts: 6
|
Tried it with RISC OS 4 (admittedly on VRPC), where PreDesk is in a slightly different place, and in which the Obey file needs a bit of tinkering. After that, I got a "Font manager in use" message. Probably doing something wrong!
Peter. Sorry, that should have been RISC OD 4.02. |
|
[ Log in to reply ] |
|
Andrew Rawnsley |
Message #118370, posted by arawnsley at 12:52, 24/7/2011, in reply to message #118368 |
R-Comp chap
Posts: 598
|
The steps look a bit odd to me, to be honest.
Since Predesk is executed alphabetically, placing files literally "in predesk" will cause them to get executed regardless, in alphabetical order. Given that ROMfonts is alphabetically after Fonts, and the obey file also loads the modules, this might be a problem.
Might I suggest making a folder called FontManager in Predesk. Inside, place Fonts and ROMfonts as per the article, and create the obey file (called !Run) in there.
Also, remember that you must reboot to trigger this. You can't just double click the !Run file in the desktop.
Oh, one last thing (which I can't quickly check, sorry). It is possible that part of the 4.02 update patches (which load in predesk - check through them) is softloading an updated font manager. It goes without saying if you're trying to use the ROOL font manager, it is the *only* one that should be loaded.
Finally, make sure you're not trying any speedup tricks like RMFaster FontManager elsewhere in predesk.
PS, a big thankyou to the author for actually producing a new iconbar article. I'd love to see iconbar have a resurgance of life - Jeffrey surely doesn't have time to keep the OS alive *and* write all the iconbar content!
[Edited by arawnsley at 20:14, 24/7/2011] |
|
[ Log in to reply ] |
|
Peter Young |
Message #118384, posted by pnyoung at 09:29, 27/7/2011, in reply to message #118370 |
Member
Posts: 6
|
[Andrew] The steps look a bit odd to me, to be honest.
Since Predesk is executed alphabetically, placing files literally "in predesk" will cause them to get executed regardless, in alphabetical order. Given that ROMfonts is alphabetically after Fonts, and the obey file also loads the modules, this might be a problem.
Might I suggest making a folder called FontManager in Predesk. Inside, place Fonts and ROMfonts as per the article, and create the obey file (called !Run) in there.
[Me] Apologies for the mangled quoting, but I can't get my head round how to do this in NetSurf. :-(
Now working up to a point, thanks. (See below for my silly mistake). I have two Unicode fonts, the standard Cyberbit and an Amharic font called Abysinica. Having got the new font manager running, I ran NetSurf; when it got as far as Cyberbit when looking at the fonts, it fell over, saying that the Unicode font manager could not be initialised. I took out Cyberbit, ran NetSurf again, and this time it worked. I could display a web page with Amharic characters. I have quite an old version of NetSurf on VRPC, so when I have time (not the next few days) I'll have another try with a newer NetSurf.
[Andrew] Also, remember that you must reboot to trigger this. You can't just double click the !Run file in the desktop.
[Me] My silly mistake; obvious when someone points it out to you!
[Andrew] Oh, one last thing (which I can't quickly check, sorry). It is possible that part of the 4.02 update patches (which load in predesk - check through them) is softloading an updated font manager. It goes without saying if you're trying to use the ROOL font manager, it is the *only* one that should be loaded.
Finally, make sure you're not trying any speedup tricks like RMFaster FontManager elsewhere in predesk.
[Me] No, it wans't softloaded here. No trick, either.
[Andrew] PS, a big thankyou to the author for actually producing a new iconbar article. I'd love to see iconbar have a resurgance of life - Jeffrey surely doesn't have time to keep the OS alive *and* write all the iconbar content!
[Me] Hear, hear! |
|
[ Log in to reply ] |
|
Sion |
Message #118386, posted by apacketofsweets at 10:50, 27/7/2011, in reply to message #118384 |
RISC OS, too cool for Javascript.
Posts: 110
|
I'd love to see the iconbar have a resurgance of life An iconbar resurgance would definitely be swell, afterall we may have the RISCOSitory and RISCOSCode but this is one of the first places to come up in a Google search for RISC OS, so a load of curious people will stumble here on occasion.
Sites like RISCOSitory and RISCOSCode are actually several pages in to a Google search, for the time being anyway.
Oh, and Peter, in order to quote a piece of text simply put [ quote] at the start then [ /quote] to finish. Alternatively you can click on 'reply and quote' and snip to your heart's content!
[Edited by apacketofsweets at 10:57, 27/7/2011] |
|
[ Log in to reply ] |
|
Trevor Johnson |
Message #118391, posted by trevj at 12:46, 27/7/2011, in reply to message #118386 |
Member
Posts: 660
|
I'd love to see the iconbar have a resurgance of life [...] one of the first places to come up in a Google search for RISC OS, so a load of curious people will stumble here on occasion. We could also do with some further references to the site's importance, for inclusion in its Wikipedia article, as its notability is under question. Any substantial references in the Acorn press would be very helpful. (Feel free to add them to the main text or just stick them on the talk page.) |
|
[ Log in to reply ] |
|
Peter Young |
Message #118420, posted by pnyoung at 14:50, 3/8/2011, in reply to message #118368 |
Member
Posts: 6
|
Tried it with RISC OS 4 (admittedly on VRPC), where PreDesk is in a slightly different place, and in which the Obey file needs a bit of tinkering. After that, I got a "Font manager in use message. Probably doing something wrong!
All working now on VRPC, thanks to help on the NetSurf mailing list, particularly Tony Moore. It seems that my Cyberbit font was probably corrupt, even though NetSurf on RISC OS 5.16 didn't object to it.
Peter. |
|
[ Log in to reply ] |
|
Tarquin Alexander Graham Mills |
Message #118613, posted by Reboot at 12:39, 9/9/2011, in reply to message #118358 |
Member
Posts: 19
|
Unicode is plot to try to get Hebrew Software on to RISC OS should be stopped, and it will be. The government probably paid for Unicode to be added. |
|
[ Log in to reply ] |
|
Rob Kendrick |
Message #118614, posted by nunfetishist at 12:47, 9/9/2011, in reply to message #118613 |
Today's phish is trout a la creme.
Posts: 524
|
No, it's a plot by the Klingons, dear. They planning on taking us over on the quiet in a change from their usual approach. Also, your imminent banning from these forums is also part of this pro-Jew/Klingon plot. They're after you! |
|
[ Log in to reply ] |
|
Sion |
Message #118615, posted by apacketofsweets at 12:47, 9/9/2011, in reply to message #118613 |
RISC OS, too cool for Javascript.
Posts: 110
|
Unicode is plot to try to get Hebrew Software on to RISC OS should be stopped, and it will be. I see you're developing your very own Operating System Mr. Mills, KnightSky OS. Will Hebrew Software be prohibited on your masterful creation? I swear there's a law against that.
Tarquin's KnightSky OS homepage: http://homepage.ntlworld.com/speccyverse/comp/saxom/index.html
[Edited by apacketofsweets at 13:16, 9/9/2011] |
|
[ Log in to reply ] |
|
Rob Kendrick |
Message #118616, posted by nunfetishist at 13:22, 9/9/2011, in reply to message #118615 |
Today's phish is trout a la creme.
Posts: 524
|
Those KnightSky OS documents are the funniest thing I've read all week. Ta, Sion. It's pretty clear from them that the author is... somewhat full of buzzwords and technical jargon but empty on fact. |
|
[ Log in to reply ] |
|
Tarquin Alexander Graham Mills |
Message #118657, posted by Reboot at 12:47, 13/9/2011, in reply to message #118616 |
Member
Posts: 19
|
That is fake part of the website created by jews, like the damage they have done to http://www.biblesabbath.org.uk is why I made http://www.biblesabbath.co.uk to give people access to the original SDA version of the website but www.archive.org (an archive of past versions of all websites) needed more of the original for copying, but jews stopped that. To see JEWS in action try these http://roshpinaproject.com/2011/03/22/paedophile-student-minister-reinvents-himself-as-messianic-rabbi/ or for fake book http://nrchurch.nr or http://www.appliedjudaism.org |
|
[ Log in to reply ] |
|
Rob Kendrick |
Message #118659, posted by nunfetishist at 13:27, 13/9/2011, in reply to message #118657 |
Today's phish is trout a la creme.
Posts: 524
|
Have you taken your pills recently, Tarquin? Why not take a nice holiday to Germany and share your views there? |
|
[ Log in to reply ] |
|
Tarquin Alexander Graham Mills |
Message #118696, posted by Reboot at 21:34, 14/9/2011, in reply to message #118659 |
Member
Posts: 19
|
Is Unicode legal on ROL RISC OS as it by Castle not ROL. I have already complained to RISC OS Open to get them to remove the ROMFonts file (Unicode font manager). The Whois information says Stephen Revill runs the website with an Ely phone code but with CJE Sussex address. Stephen obviously does not exist and for Unicode Castle aRe Evil.In office hours CJE answer, the rest of the time Castles answerphone is used. RISC OS Open sounds like a plan to open source (GPL) RISC OS, already they have put a binary copies of several OS parts on the website. The phone number is 01353740267 but if ring they ring the police.
[Edited by Reboot at 09:47, 16/9/2011] |
|
[ Log in to reply ] |
|
Philip Webster |
Message #118697, posted by pwx at 21:55, 14/9/2011, in reply to message #118696 |
Member
Posts: 227
|
Is Unicode legal on ROL RISC OS as it by Castle not ROL. I have already complained to RISC OS Open to get them to remove the ROMFonts file (Unicode font manager). The Whois information says Stephen Revill runs the website with an Ely phone code but with CJE Sussex address. Stephen obviously does not exist and for Unicode Castle aRe Evil.In office hours CJE answer, the rest of the time Castles answerphone is used. RISC OS Open sounds like a plan to open source (GPL) RISC OS, already they have put a binary copies of several OS parts on the website. This is a *brilliant* bit of trolling, but I'm not going to bite. |
|
[ Log in to reply ] |
|
Sion |
Message #118700, posted by apacketofsweets at 23:46, 14/9/2011, in reply to message #118696 |
RISC OS, too cool for Javascript.
Posts: 110
|
Is Unicode legal on ROL RISC OS as it by Castle not ROL. I have already complained to RISC OS Open to get them to remove the ROMFonts file (Unicode font manager). The Whois information says Stephen Revill runs the website with an Ely phone code but with CJE Sussex address. Stephen obviously does not exist and for Unicode Castle aRe Evil.In office hours CJE answer, the rest of the time Castles answerphone is used. RISC OS Open sounds like a plan to open source (GPL) RISC OS, already they have put a binary copies of several OS parts on the website. If I didn't know better I'd say Tarquin could be suffering from schizophrenia, either that or he hasn't discovered women yet. I know what one I'm going with. |
|
[ Log in to reply ] |
|
Tarquin Alexander Graham Mills |
Message #118916, posted by Reboot at 15:53, 2/10/2011, in reply to message #118700 |
Member
Posts: 19
|
Netsurf has a Boot Resource called !Unicode that has Hebrew files for Hebrew fonts. If you delete !Unicode Netsurf stops working. The icon has a kanji charcter rather than tell the truth and having a Hebrew one. |
|
[ Log in to reply ] |
|
|
The Icon Bar: News and features: Getting Unicode Working With RISC OS 4 (and 6?) |