Andrew Grant: There are at least three schemes for romanization of Japanese, but SL uses (and should use) the "wapuro" style, which avoids diacritical marks, as the use of special characters (vowels with macrons) makes the search functions impractical to use.
Bob McGuigan: I agree.
DrStraw: I think there are three considerations:
- Search capability: It is essential that people can find references to terms using the search utility, especially the important ones like Honinbo.
- Ease of user: People need to be able to enter the term from a standard keyboard. Use of diacritical marks should be discouraged unless they are essential for understanding.
- Comforming to standards where available: For example, "Honinbo" has been standard and is used by the Nihon Kiin.
Of these three the first is by far the most important. As Japanese is the only Asian language of which I have any knowledge I cannot comment on others, but a I would assume that the same considerations would apply.
Andrew Grant: A related question is whether we should indicate long vowels. The wapuro style allows this; for instance "Honinbo" could be "Honinbou". Now, I'm not suggesting we do this for words like Honinbo which have become standardised, but maybe for less common words we could consider it?
Alex: I think consistency is important, so if the Japanese are leaning towards not indicating long vowels (is this true, or is Honinbo an exception?) we should follow suit. I do think that long vowels should be indicated on the word's main page (e.g. Note that the final O is long, i.e. "Honinbou," although "Honinbo" is the standard romanisation).
Andrew Grant: The Japanese are not leaning towards not indicating long vowels; it is purely a Western habit. (Not indicating long vowels can cause serious problem in Japanese; consider the difference between komon which means adviser, and koumon which means anus!)
(Dave: The urge to follow up on this example with a few lawyer jokes is almost unbearable!!! :-)
Bob McGuigan: It is possible to use unicode in the search field. However the user then has to have the knowledge of how and when to do that. I wonder how many SL users have this knowledge. Moreover, due to a long practice by Go World magazine, Ishi Press, and Kiseido of ignoring the "long" vowels in Japanese, most Western go players probably don't know when "o" and "u" are long. To add to the problem, sometimes in wapuro romanization a long "o" is written "oo", as in ooba rather than "ouba".
kokiri: I think i agree with pretty much everything above. I would suggest that in the main we stick with what we have used up to now - generally not introducing n' or indicating long vowels. I think that the n' is elegant, but adding it would either have to be inconsistant, or alter the standard form for some words. My personal preference would be to ignore long vowels because 1) i think they are less attractive & 2) they end up being less helpful for pronounciation for those who don't know any japanese in addition to the fact that they are not usual in a Go context. Maybe the thing to do is to have Honinbo as the standard usage, but to say on the main page, Honinbo, perhaps more accurately romanised as Hon'inbou for reference. in some situations you also get oh indicating a long vowel, to add to the confusion
Andrew Grant: If we're going to ignore long vowels, someone should go through SL to ensure it's done consistently. A year or so ago, when a number of people were learning Go after reading the online translation of Hikaru no Go, we got a lot of newbies here using words like "shidougo" and "Shuusaku" because the HnG translators were using wapuro long vowels.
Velobici: Regarding shidougo and Shuusaku used by those that come to SL from Hikaru no Go, lets provide for them by creating aliases: shidougo -> shidogo and Shuusaku -> Shusaku. This way we both welcome them and introduce them to standardized spellings.
unkx80: I think this is a good idea. Perhaps, also list the alternative romanizations on the page defining the terms?
viciousman: I think that a modified form of Revised Hepburn (without macrons or apostrophes) should be used for the words. For when the Japanese characters are shown and a more accurate romanization is needed, I think the actual Revised Hepburn should be used in that case, since computers that cannot handle unicode (and therefore would require a macron to be a French circumflex instead) are becoming increasingly rare.
Bob McGuigan: The following is transferred from my home page:
viciousman: Regarding this comment: "Using ou for the long o is no less standard romanization than revised hepburn so there is no need to change everything to revized hepburn. In fact, doing so messes up the search function. Using the diacritical marks requires special unicode characters, another reason not to use revised hepburn." The reason why Revised Hepburn is considered to be a standard is because the U.S. Library of Congress adopted it and most academic sources (e.g. encyclopedias) now use Revised Hepburn.
Bob McGuigan: One more piece of information to consider is that some major online dictionaries do not accept revised hepburn using unicode. In particular Kanjidict 1.7.0 accepts 急所 (unicode) as input and gives kyuusho as the romanization but seems not to accept "kyūsho" as romanized input.
As I mentioned earlier by far the major part of English go literature ignores the "long" vowels. Whether this is a good or bad thing is beside the point since the books and magazines are there and aren't likely to change. Thus anyone searching on SL for a term or name from Go World, an Ishi Press book or a Kiseido (not Kiseidou) book would not use long vowels in any form. This means we have to have aliases for these terms without the long vowels whatever else we do. In fact this short-vowel-only practice essentially creates English versions of the Japanese words. This sort of thing is not uncommon with English and other languages, for example romanization of Russian words and names, or even "English" versions of French, Spanish, or German words.
viciousman: See if the paper dictionaries are any different. The online dictionaries are set up that way because most English-language keyboards do not have macrons on them! Also see if the online dictionary accepts the French circumflex. However, I do not mind having words usually as like Hepburn but without long vowels or apostrophes. I would use Revised Hepburn whenever showing a more detailed romanization (e.g. in italics next to Japanese characters).
iopq: Wow... that's a stupid way to Romanize. I didn't even know that it was really hon'inbou! I've been pronouncing it wrong all my life and it's the fault of people like you. No, wait, my major source of Go information is SL so it IS your fault.
Just use Hepburn. It works. Period.
"Romanization of Chinese is difficult."
Malweth Unlike in Japanese, there are many different ways of romanizing Chinese. "Qing" and "Ching" are phonetically equivalent. Qi is a Chi sound when using the Chinese romanization based on Russian languages.
spindizzy1976 I'll make the disclaimer that my Chinese is pretty rubbish... Still, I think it's a bit of a simplification to say that "qing" and "ching" are phonetically equivalent. There is no "ching" in Mandarin Chinese although maybe it exists in some of the dialects, but anyway, "qi" and "chi" are audibly different sounds even to non-Chinese speakers.
I prefer pinyin to use tone marks 1-5 after each character e.g. pin1yin1. That method is fairly common on computers and doesn't need any technical wizardry to type in. More importantly (IMO), I notice there are many places on SL where only simplified characters are given. I think it would be nice (and respectful to those outside mainland China) if the traditional forms could be provided as well.
unkx80: Unfortunately, using diacritical marks is the official standard. As the maintainer of Chinese Go terms, I definitely know that entering diacritical marks is such a pain in the ass, but I get repeated requests for such, and so here they are. Luckily, there are also standard rules on where to put the diacritical marks, so I simply enter tone numbers in my database and write a program to convert them to diacritical marks. =D
As for traditional Chinese characters, they are also on Chinese Go terms by popular request.
tderz: Mainland China propagates Pinyin (as the one and only romanization) since about 25 years.
Example: Mao Zedong (毛主席, Pinyin: Máo Zédong1) is not written anymore Mao Tse-Tung (Wade-Giles). In Taiwan Zhuyin Fuhao is used. 
unkx80: The Chinese characters you gave is Mao2 zhu3xi2, which means Chairman Mao.
I think the complicatedness of a language or its transcription into another language does not matter so much. BTW, pinyin is easy and straightforward, IMO. Some letters get an unusual pronunciation (e.g. "x", "q", "r" a.o.) w.r.t. western speakers, but the tone differences were actually more important in speech.
Because I only know Pinyin (a bit) I cannot really judge the others, but Zhuyin Fuhao ( Zhuyin, Tongyong_Pinyin) uses again other characters and does not help very much for typing it with Western letters, and Wage-Giles is very confusing to me.
Two factors are very important:
You smell it already, for me there would be only one romanization system: Pinyin
Very, very theoretically, Chinese characters could even be replaced by pinyin, this is (has been) discussed in China.
I think this would be a pity - after all, the written language is the same for many people who could not communicate orally together.
unkx80: While I personally use Pinyin (because it was the one taught in school), and do not know the other romanization systems well, I think Wade-Giles uses a romanization that is more consistent with the way Europeans pronounce letter sequences, compared to Pinyin. I suppose I am now used to hearing Europeans pronouncing Pinyin names in all kinds of weird ways... All the above is just for your information only. Anyway, we all agree to use Pinyin on SL nonetheless.
 Phonetic variants of "Deng Xiaoping"