Text 6936, 134 rader
Skriven 2013-12-10 15:08:22 av Nicholas Boel (1:154/701)
   Kommentar till text 6902 av mark lewis (1:3634/12.71)
Ärende: UTF-8
=============
Hello mark,

09 Dec 13 20:28, you wrote to me:

 NB>> I think everyone should be able to write their names exactly how
 NB>> they are spelled. I'm not sure as to why this is such a huge
 NB>> argument on Fidonet, but UTF-8, Latin-1 (8859-1), koi8-r, and
 NB>> every other widely used charset out there should have been dealt
 NB>> with when Fidonet became an international network.

 ml> that was back in the '80s and early '90s... UTF8 was just getting
 ml> started around '92 and didn't get really going until Jan '93 when it
 ml> was first presented at the USENIX conference... it wasn't until 2008
 ml> when utf-8 was recognised as the most common encoding for html
 ml> files...

Then now in 2013, most linux distrobutions are using it as the default locale
out of the box. While we do not have heaping quanities of newcomers to BBSing
and/or Fidonet these days, which probably has some reason as to why we're not
seeing a lot more systems show up posting in UTF-8. It is, in fact, becoming
the go-to international charset. Obviously M$ will always try to create their
own, since it wants all of the marketshare, but it seems linux/unix are all
moving in that direction. Fidonet is behind, mainly because of the continued
use of old software that will never be able to include support for what the
future will bring. If/when this time does come, people will have an difficult
choice: update their software, or shut it down for good. I would hope people
would choose the first option, but if not, many will be moving on to other
things.

 NB>> The rules of English speaking only echos can still hold their
 NB>> weight, but no charset or codepage has ever been deemed the
 NB>> *only* one you are allowed to use.

 ml> because the developers at the time saw the need for being able to
 ml> translate into other character sets to enable others to communicate
 ml> easier... unfortunately they weren't as successful as they could have
 ml> been and they left fidonet for various reasons before they could
 ml> complete their tasks with their software... many left in disgust due
 ml> to others actions and posts which we still see a lot of similar type
 ml> stuff today... some few remain but they don't release their code or
 ml> talk about their work because of the way others talk to and about
 ml> them...

The translations are great, don't get me wrong. But for some reason, my GoldEd
cp866 -> utf-8 and koi8-r -> utf-8 translation tables that come with it don't
seem to work very well when it comes to Cyrillic. Maybe some of those
characters weren't translated at all, or it could even be possible that some
people are throwing the wrong CHRS kludge on a message without knowing. Could
also be because GoldEd doesn't currently support UTF-8 unless you use the odd
'workaround'. I don't know. I do know that I can open any raw packet in nano
and so far have been able to see most of what others and I have thrown at it
exactly how it's supposed to be seen. That leads me to believe it's the editor.
:)

 NB>> It has been said in the past, people that don't care to upgrade
 NB>> or re-configure their software to this day and age will see
 NB>> garbage on their screens. So what? It doesn't break anything on
 NB>> their end, except their visual experience.

 ml> and that's over half of the reason to participate in the network... if
 ml> everyone i converse with were to switch to using utf-8 and non-ASCII
 ml> characters, i would probably leave, too... unless i was able to
 ml> develop something that fit my needs... whether i do that or not
 ml> remains to be seen but i do have several megs of code that i have
 ml> written which is sitting idle and has been for years... some i used to
 ml> use during development with an eye to releasing it for others to use
 ml> but the abuse took that idea away and lead to a quite angry and
 ml> abusive life due to others actions...

This wasn't really directed at using utf-8 and non-ascii characters in messages
to the point noone would understand it. This was more of a discussion on people
being able to write their names how they should be written. Most of what you
see in regular messages will still be readable.

Then again, the same thing could be said about "high ascii" characters being
used. When one types them in ibmpc or cp437 they need to be translated to
latin-1 or utf-8. I don't even know of any current OSs that use ibmpc or cp437
anymore. Maybe OS/2?

 NB>> That is not the fault of people wanting to keep their software in
 NB>> the current century in regards to international keymaps,
 NB>> charsets, and codepages. *shrug*

 ml> the problem is that even though you may be able to properly see eg:
 ml> cryllic or greek or chinese glyphs, it doesn't really help you read
 ml> and participate in the conversation... at best you might be able to
 ml> paste it into a translator and get something that might or might not
 ml> make sense in english...

First thing is first. You need to be able to see it properly to even try to
translate or understand it. You have to start somewhere, and as I mentioned
before, if I could see it properly, I may actually be interested in carrying
some popular echos here just for the simple fact of venturing into the rest of
the world.

 ml> anecdote: i had to really laugh earlier at another's utf-8 postings
 ml> where their russian version offered several variations of 'protein'
 ml> first and finally offered 'squirrel' as an option... really?

 ml>   "i have protein in my pants"
 ml> OR
 ml>   "i have a squirrel in my pants"

 ml> :LOLOLOL:

 ml> at least that was better than that translator that rhymes with 'thing'
 ml> ;)

The only real translators you can trust are real people that are fluent in both
languages. I'm sure there's different meanings in the way you say things (just
like English) that computers will never understand. At least you can get a
general idea of what someone is trying to say with the online translators, even
though often it may be completely wrong at the same time. :)

 ml> agreed to a point... did you know that what many folk consider to be
 ml> "modern" software (ie: internet apps) really use '60s tech for their
 ml> transmissions? eg: attaching a binary to an email... the binary is
 ml> translated into plain ascii and gains generally 1/3rd the size it
 ml> would be without the translation (MIME, UUE, XXE) ;)

Is that because it's easier to use a library or point your code at something
that's already in existance that has worked for many years, rather than doing
it yourself? If so, you can just chalk that up to lazy coding practice, I
suppose. I'm sure careered people are usually on strict deadlines, so why write
a new library, or make something from scratch, when something is already there
and easy to implement, which will save you time so you can grab a beer the
night before turning in your project, rather than cramming all night to get
that project done before work the next day? :)

Regards,
Nick

--- GoldED+/LNX 1.1.5-b20130910
 * Origin: Dark Sorrow | darksorrow.us (1:154/701)