Learning 五笔字型 for traditional Chinese

[shibole]

(I thought maybe this subject was unique enough to deserve its own thread. Please merge it to do whatever if not.)

I'm currently learning traditional characters. Imron was saying that I should learn WuBi 18030 (not sure where that number comes from) because it supports traditional Chinese. Some questions:

From what I've read, WuBi is better (easier to learn, more consistent, etc) than other similar input methods like cangjie, but I tend to associate things like cangjie with traditional. Am I doing something "unnatural" that people don't normally do by trying to use WuBi for traditional? (Like does everyone in Taiwan/HK or something use cangjie, and nobody anywhere uses WuBi for traditional?)

五笔快打 says that I need to downgrade to v5.1 to learn 18030. From using v6.1 so far it looks like there's no "traditional" mode that would give me practice with traditional characters. Does anyone know if I should just downgrade to solve all my problems, or is 五笔快打 maybe not the more suitable thing for learning wubi with traditional?

I've also noticed that some things (like inputking) seem to have a "Traditional" mode for wubi86. Does this mean 18030 is only needed if you want some "modeless" way of typing both with the same input system?

[imron]

18030 comes from GB 18030, an encoding standard for Chinese characters that is the successor to the GB 2312 standard. GB 2312 was back in the days before unicode and only had support for ~6,000 characters, most (all?) were simplified. GB 18030 encompasses the entire unicode code space (and is defined relative to it), and therefore has support for traditional characters. Wubi 18030 is the version of Wubi that first included support for typing these extra (and hence traditional) characters.

Anyway, yes, you are correct in that most people who use stroke-based input methods for Traditional characters will use Cangjie and not Wubi. Most likely because Wubi was invented in mainland China and for a long time (more than a decade) only supported simplified characters.

As for versions, I'm still using 五笔快打 4.1 and have no problems with it. If the latest version doesn't have a traditional mode the other thing you could do, assuming you were already familiar with which keys map to which shapes, is create your own practice articles with Traditional Characters, and use them. I find using my own practice articles is the key to improving speed anyway, as I can make sure the content is all characters that I know the meaning of.

I'm not sure if they changed the interface much in the newer version, but in my version, the option to create your own articles is under: 打字练习 -> 自选练习

[shibole]

I'm not sure if they changed the interface much in the newer version, but in my version, the option to create your own articles is under: 打字练习 -> 自选练习

Actually they seem to have added a huge area of functionality in 6.1. Specifically this sort of thing:

Basically this menu doesn't exist in 5.1:

Also you can isolate your practice to certain keyboard regions which can make learning what "roots" are on what key a little less frustrating:

So basically I was hoping that this part of the software would support traditional characters, but it doesn't. Maybe I'll just use this to get used to what roots are assigned to what keys and then switch back to 5.1 and practice with my own input text or something.

It seems like 6.1 supports the 18030 dictionary but the "hint" thing over on the right can't be configured to display 18030 hints. Oddly though it seems like it might be displaying 18030 hints anyway because I have the 18030 dictionary selected?

For example, if I create a file with 醫生 and 医生 then I get ATDG for 醫 and ATGI for 医 in 五笔快打 even though supposedly I'm using wubi86. 极点五笔 (free wubi IME) won't seem to let me type both, at lest the way I have it configured. I'll have to mess with it.

But anyway you see why I'm a bit confused here. I'm not sure what the difference is between wubi86 with a 18030 dictionary and wubi18030.

And when am I going to be able to type biáng?

[ipsi()]

AppLocale doesn't work well for 五笔快打, do I need to set my system language to Chinese (Simplified) for it to work properly?

[shibole]

Yes

[ipsi()]

Figures. Oh well. Thanks for that.

[imron]

Strict wubi86 doesn't support traditional characters, and is so named based on the year it was created (1986). Wubi 18030 is named based on the character encoding that it supports, and uses the same character roots as wubi86 with some added new ones to support traditional characters. This is different from say wubi98 which actually changed some of the roots onto different keys. Due wubi98's lack of popularity, when the company that invented Wubi released Wubi18030, they went back to wubi86 key mappings. So, wubi 18030 is essentially wubi86 with traditional character support. To reiterate, most traditional characters could theoretically be typed using wubi86, it's just that the encoding used at the time only had support for ~6,000 characters, which were mostly (all?) simplified ones, and so traditional characters couldn't be typed out. Wubi 18030 rectified that, and by using the newer GB18030 encoding, it has support for all characters (both simplified and traditional).

Next, I find it strange to think that Shunsoftware would have removed functionality from an older version. Are you sure the older version actually does what you want?

Also, to be honest, if you are just wanting this functionality to learn the key mappings, I don't think it makes much of a difference if you are learning traditional or simplified. 18030 and 86 are almost identical in this respect. This is because the character roots are usually smaller than a radical, and traditional and simplified characters are essentially made up of the same core shapes - it's just that the placement of these shapes is different. Wubi 18030 has a couple of extra additions for things like 車 and 門, but these shapes are typically placed on the same keys as the simplified variant or on other easily guessable keys (easily guessable based on the way the "5 pens" are laid out). Once you've got a feeling for what shapes go on what keys, then you can start to use the other training sections, because the only real difference will be in how you combine the shapes. A good example is the one you listed above for 医/醫. The character 醫 (like the vast majority of traditional characters) isn't made up of any roots that don't exist in wubi86, so once you've learnt the wubi86 key mappings, you'll be able to type most traditional characters anyway. The only difference is the occasional shape like 門 or 車 mentioned above, but these are the exception rather than the rule and again the only real difference is how you combine the roots, e.g. simplified 开 is GA, compared to traditional 開 UGA.

Finally, to type 醫 using freewb, you should click on little icon on the freewb toolbar that says "GB". This should now change to say "GBK". Now you can type ATDG to get 醫, or ATDI to get 医. (note, it should be ATDI for 医 and not ATGI).

[ipsi()]

Where does one get a Windows IME for Wubi 18030? I've got one for 86, which is all good, but I'd like to have the option of typing Traditional as well.

Also, a couple of questions about 五笔快打: How do I access the other sets of exercises? I can currently do the ones for the asdfg keys, but that's it... Also, given that I'm not in China, how do I go about buying i

[shibole]

Next, I find it strange to think that Shunsoftware would have removed functionality from an older version. Are you sure the older version actually does what you want?

I installed both of them and looked at each. The older version (5.1) does actually have options for 18030 that 6.1 lacks, but 5.1 totally lacks all the new exercises that teach character decomposition and key assignment. It sort of makes sense for them to "remove" 18030 because they probably didn't want to have to create the new character decomposition exercises for all of 86, 98, and 18030 in their first release of the feature. Seems like they could have left it in for use with the older functionality of just typing the articles though....

You license key may actually work for the newer ones, so if you want to play with them you might be able to do so. (My "6.1" license key worked for 5.1. No idea if v4 uses the same licensing system though.)

6.1 has some menu buttons at the top of the window that say wubi98 and wubi18030. When you click on one, a little box appears explaining how you need to download v5.1 for 18030. I can't read the whole thing so I've attached an image.

Also, to be honest, if you are just wanting this functionality to learn the key mappings, I don't think it makes much of a difference if you are learning traditional or simplified.

I was pretty sure that it would be perfectly fine for learning key mappings, but the other thing it does (if you check out the other exercise types) is teach you the key mappings in the context of actual characters (shows a character, highlights the part corresponding to the key, then tests you on your ability to find the key). Again this is probably still pretty much the same between traditional and simplified, if I had to guess anyway.

However the thing that looks like it should be really useful is that it will teach you individual characters root-by-root, highlighting each one. So it basically teaches you the "logic" behind each key sequence for some number of characters rather than just telling you what the key sqeuence is. Still, the logic may be close enough between traditional and simplified but...

Finally, to type 醫 using freewb, you should click on little icon on the freewb toolbar that says "GB". This should now change to say "GBK". Now you can type ATDG to get 醫, or ATDI to get 医. (note, it should be ATDI for 医 and not ATGI).

Odd, the default theme didn't have that GB/GBK button. I fixed that.

(Yea, sorry I gave the wrong code for 医. I'm still pretty confused about this stuff.)

For 醫 for example I understand the AT but I don't understand the logic of D here at all. I think maybe the G is just the first horizontal stroke on the 酉? Anyway this is why I was hoping 五筆快打 would teach decomposition of traditional characters.

Do you happen to know of some sort of wubi dictionary out there that shows actual decomposition rather than just the code?

Where does one get a Windows IME for Wubi 18030? I've got one for 86, which is all good, but I'd like to have the option of typing Traditional as well.

I thought this one did but I only see 86 mentioned: http://www.freewb.org/

I was able to type 醫 with it though ATDG resulted in 2 matches and I had to pick the second one.

Also, a couple of questions about 五笔快打: How do I access the other sets of exercises? I can currently do the ones for the asdfg keys, but that's it... Also, given that I'm not in China, how do I go about buying i

You have to register it or you can only use region 1 in the exercises.

You need a Chinese bank card to register it. I had to get my brother-in-law to do it for me, then it took me forever to get the code because I had him give them my email address and somehow it didn't work. If you get someone else to register it for you I strongly recommend that they give their email address then forward you the code. It took my brother-in-law several calls to get ahold of someone to have the code sent to his email address after he had originally given them mine.

[ipsi()]

Thanks for that. I'll have to see if I can get someone in China to buy it for me. At 25RMB, that's pretty damn cheap.

[shibole]

At 25RMB, that's pretty damn cheap.

Yea, which is a big reason I don't mind buying it even if it doesn't do everything I want perfectly.

[ipsi()]

Indeed. Probably explain why I can't find it via other sources. Meh. I do wish it was a bit easier to purchase though.

[roddy]

The reg.banma.com site where I tried to buy it from now has an option to pay via Shenzhouxing card, if that's of any use to anyone.

[imron]

The dialog just says you need version 6.1 (Wubi 98 edition) if you want to learn Wubi 98, version 6.1 (Wubi 86 edition) if you want to learn Wubi 86, or version 5.1 if you want to learn 18030, and then mentions that the serial numbers for version 5, 6 and 6.1 are all interchangeable.

Version 4 uses a different licensing system that is not compatible with the newer versions. I was thinking about upgrading (at the time it would have been a free upgrade), but the newer versions didn't seem to add anything I needed, and when they didn't reply to an email enquiry I never bothered to follow up on it .

So it basically teaches you the "logic" behind each key sequence for some number of characters rather than just telling you what the key sqeuence is. Still, the logic may be close enough between traditional and simplified but...

The logic is close enough that you can interchange it. Once you know the logic, you know the logic and can apply it to any given character you see. The other thing to remember is that a reasonable amount of simplified and traditional characters are the same (or mostly the same) anyway, so it's not necessarily wasting your time.

For 醫 for example I understand the AT but I don't understand the logic of D here at all. I think maybe the G is just the first horizontal stroke on the 酉? Anyway this is why I was hoping 五筆快打 would teach decomposition of traditional characters.

Because this character has many roots, the overall logic will be: Press the key for the first root, press the key for the second root, press the key for the third root, press the key for the final root.

Here is a breakdown of all the wubi roots in the character:

The first root is 匚 - A.

The second root is  - T

The third root is 大 - D (T and D put together gets you 矢)

The fourth root is 几 - M

The fifth root is 又 - C

The sixth root is 西 - S

and the final root is 一 G.

Why is the last root 一? Well, because there is no root for 酉. The S key only has the 西 root. This means that you still need to add a 一 to complete the character, thereby making G 一 the final root. Combined with what I mentioned above about typing the 1st, 2nd, 3rd and final root, it should now be clear why the final code is ATDG

This further illustrates what you were talking about by the logic. I never really type Traditional characters, and I only know the logic of Wubi from practicing Simplified. However should I wish to type Traditional, that logic is directly applicable.

Do you happen to know of some sort of wubi dictionary out there that shows actual decomposition rather than just the code?

Nope, this is just one of the things that comes with practice.

I thought this one did but I only see 86 mentioned: http://www.freewb.org/

I was able to type 醫 with it though ATDG resulted in 2 matches and I had to pick the second one.

This is what I'm using, and can type in Traditional just fine. I think it's just the naming terminology is just confusing. Just think that any modern Wubi86 (i.e. one that can type traditional characters) is the same as 18030 (they use essentially the same key mappings) and you'll be fine. Also for me, using this IME, 醫 is the first character choice for ATDG. One thing you might want to do is go into the options and under the 基本设置 make sure that 启用自动调频 is turned off. This will stop the IME from automatically reordering the options when there is a keycode "clash".

[shibole]

The first root is 匚 - A.

The second root is  - T

The third root is 大 - D (T and D put together gets you 矢)

Ok, that makes more sense. I was thinking perhaps that the 2nd root was 禾 or something, since I've encountered some cases where some vaguely similar character like 六 is considered to be something you wouldn't immediately expect like 大. Hopefully if the decomposition rules in 86 are consistent enough then I'll just pick all this up regardless of character set....

I think it's just the naming terminology is just confusing. Just think that any modern Wubi86 (i.e. one that can type traditional characters) is the same as 18030 (they use essentially the same key mappings) and you'll be fine.

I'm not so sure that's the case. Based on this from the FAQ at wbfans.com (http://www.wbfans.net/Article/ArticleShow.asp?ArticleID=60):

18030版 — 王码根据GB18030-2000大字集标准推出的针对第一代五笔字型的加强型版本，编码规范性做了改良，通过了国标鉴定，但受86版和98版用户影响，目前其使用人数是最少的。这三个版本无任何速度上的差别，速度主要取决于练习方法和练习强度。

So I don't know if the improved "编码规范" makes it easier to learn or dramatically improves the logic of character decomposition or what... maybe I'll try to find out the specifics, but maybe it doesn't really matter that much. As long as I'm not totally locking myself into simplified I guess I don't care since I just want some way of typing that helps me remember how to write the characters that is easier to learn and more logical than cangjie. (Cangjie looks hard even for people who are extremely familiar with the language, which isn't too surprising as it was originally invented in 1976.)

[imron]

So I don't know if the improved "编码规范" makes it easier to learn or dramatically improves the logic of character decomposition or what.

Nope, the logic and learning curve are the same. It just means the range of characters has been increased. i.e. before wubi could handle 6000 characters, now it can handle 27,000 (or something like that). The core concepts are still the same, it's just that the word database has been increased. i.e. in wubi86 there is simply no way to type 醫, and typing ATDG would result in no matches. Now, the code ATDG (along with a whole bunch more) have been added to the list of codes and so typing it will produce a result. Load up 五笔快打, and open up the keyboard maps and spend a bit of time comparing the 18030 and 86 keyboards. You will see that they are more or less identical except for the addition of roots to support the extra characters. Wubi 98 on the other hand has moved some of the roots around.