English Chess Forum

Welcome back the real O'Gorman! Well done Carl, leading candidate for the Apostrophe Lovers Association award of man of the year.

Brendan O'Gorman wrote:Welcome back the real O'Gorman! Well done Carl, leading candidate for the Apostrophe Lovers Association award of man of the year.

The Russian posts and some of the other character issues will have to wait until after the holiday everyone.

New posts are fine just need to convert the latin posts that remain.

Carl,

Is a suboptimal approach worth considering? For instance doing a find and replace on, say, "émigré" and similar. Was that how you solved the "apostrophe" problem? It should be possible for forumites to come up with a list to be converted.

I can only recall two non-Latin scripts being used: Russian and Hebrew (transcribing, "Ze lo democracy" in the latter springs to mind). My belief is that only the former is currently significant from the perspective of the EC Forum. My impression is that most of the Russian text can be found within the Forum's quotation markers, i.e. [quote][/quote]. If the problem is finding a bijection, then why not reverse map the text found within such markers? If the resultant values are all within the Cyrillic range for UTF (From memory that's 0x0400—0x04FF, my eyesight is too screwed up at the moment to read numbers accurately. That is also why it would be difficult for me to correct by inspection.) then one has a domain to map from. There may be some English text in a tiny minority of the Russian quotes. If the "all" criterion works, then perhaps a "90% test" will filch out this balance.

It may be a time consuming hack, but if the number of "Russian" posts changed is fewer than a hundred, it has probably worked.

I think it is any text containing diacritics. There are examples of "née" that are messed up. Such as here:

http://www.ecforum.org.uk/viewtopic.php?f=27&t=2165

I'm not particularly fussed about that. Agree that the Russian ones could/should be fixed as Simon suggests.

Christopher Kreuzer wrote:I think it is any text containing diacritics. There are examples of "née" that are messed up. Such as here:

http://www.ecforum.org.uk/viewtopic.php?f=27&t=2165

I'm not particularly fussed about that. Agree that the Russian ones could/should be fixed as Simon suggests.

The problem is quite widespread so building enthusiasm still here.

Carl Hibbard wrote:

Christopher Kreuzer wrote:I think it is any text containing diacritics. There are examples of "née" that are messed up. Such as here:

http://www.ecforum.org.uk/viewtopic.php?f=27&t=2165

I'm not particularly fussed about that. Agree that the Russian ones could/should be fixed as Simon suggests.

The problem is quite widespread so building enthusiasm still here.

Does that mean if we wait for enthusiasm to arrive, it will magically get fixed without the need for a clunky manual solution?

Maybe but nice weather and a beer garden takes priority even over this.

Bizarrely, I've just read this having spent most of this morning sorting out a problem with one of my websites where an inconsiderate user had the cheek to use a backtick (as opposed to an apostrophe) in his surname and broke the user management plugin code. Beer and sun definitely takes priority

Something like CONVERT(CAST(CONVERT(fieldname USING latin1) AS binary) USING utf8) will probably do the trick but not after beer.

Carl Hibbard wrote:Something like CONVERT(CAST(CONVERT(fieldname USING latin1) AS binary) USING utf8) will probably do the trick but not after beer.

Alignment could be a pain.

I can't see consecutive letters at all well, probably not for at least another month; nonetheless, I am prepared to go through the Cyrillic texts, for the majority I almost certainly have the originals or links to a suitable source. I would, however, require temporary administrator privileges when I am ready, probably not for at least six weeks. Note, as well, that I would also rationalise threads such as the two devoted to Bohatirchuk, I would merge them and clean up the neo-Nazi pollution (i.e. praising the Waffen-SS to the nearest available kike, insisting that a fake from the fifties is a "Soviet source" and such).

Before beginning, I would need Carl's assent. Is it forthcoming?