true international hub support / Unicode

[bomk]

Right now, every language area has its own hubs with every file sharing system. You cannot look for a file that even partially contains Japanese characters from a German OS, for example. Take anime: There is a huge english speaking community which still would be happy to get more of it, and even if a filename contains Japanese characters. Just one word to describe an anime file in English next to the Japanese filename would suffice to recognize it. But right now this is impossible. You have to have a Japanese system to look for such files. And if you have it you cannot look for French files.

And so on. It's the same thing with Russian, Turkish, Chinese Traditional, Chinese Simplified (mutually exclusive once again!), German. The old codepage dilemma.

Thus I propose, as a long term strategy, to add full Unicode support to DC++.

The vision is that communication between the clients would be switched to UTF-8 if necessary, and that the client itself would be fully Unicode aware so it could e.g. display a Japanese filename on a German system. As the Windows world is moving away from Win 9x systems this has become possible.

[RobbeZ]

seems like a good idea 2 me

[LordAdmiral]

Yeah, with this, maybe we could put the dollar sign back in (it'll just get sent as a different, probably unused or a similar symbol, and interpreted back as the dollar sign)... :)

The only thing I can see troubling is compatibility. I'm not sure if NMDC supports unicode, and if people start sending unicode characters into a hub chat that's running the NMDC hub software, it might crash a thing or two.

[LinkSync]

I hate that argument!
Sorry
But I just do!
Its holding back to much inovation and upgrades.
If we was still compatable with dinosoaurs we would be running around killing each other...
Well
Anyway u know what I'm getting at i bet
At somepoint both the hubware and the clientware will both need a re-write to enable many of the more modern and useful user friendly features thought of.
When does evolution become revolution?
When u dont want the change...
Change is inevitable and GOOD!
Why worry for the dinasours when we should be working to PROMOTE homoerectile!

suggestion for a good nic:
Rumpled Forskin

peace

[bomk]

The compatibility argument is used as the legendary KillItAll.

How about using UTF-8?

There would not be a big difference to the status quo . Old software would work, but display some characters incorrectly. Which happens even now because of the codepage problem.

Have made that suggestion before, at other places. Have to confess that I'd be surprised if a serious coder would actually consider it. Seems like every single programmer all the time and infallably only uses his own mother tongue, and maybe the most basic of English.

[sandos]

I dont really see the problem with this. Encode away, and dont care if it comes out wrong. Either

1) your in a hub where most people are using the software you are, maybe even recommended/required by the hub
2) your in minority, and may be kicked for writing ugly stuff.

Preferrably use the html-ized way that DC++ now does with | and $:

http://dcplusplus.sourceforge.net/forum/viewtopic.php?t=1259

[bomk]

unfortunately its not that easy.

For example I like Japanese anime, and sometimes I come across a file that looks like this "$%§&/&%/&!blah45&$§$". That might be a file with a Japanese filename, but it might also be one with a Chinese filename. Because DC++ doesnt use Unicode, I cant read it, not even verify which language it is.

The same thing happens with a French file, if you look at it from a DC++ running on a Japanese system.

How to solve this - use UTF8 for the search protocol, and convert string handling within DC++ to unicode. And let it run on WinNT sp5 or higher, not Win9x.

Unfortunately I can hack, but not program. So I cant do this, and judging by experience I doubt any serious coder will consider this kind of proposal. Thats my story.