Database
Fredrik Ekman
ekman at lysator.liu.se
Thu Oct 20 19:13:54 CET 1994
Here's yet another exciting message about database format. You have been
warned!
Right, so there are three of us at least. Harry: You managed to
straighten out a few question marks for me, but it also seems that you
misunderstood me on a few points, no doubt due to me being unclear as
usual.
Harry:
> You can explicitly indicate in an 'issue' file that a story
> has a title, differing from the "official" one.
Yes, but can you see in the .issues file if the title is for this issue
specifically or "official"? I don't think so (although correct me if I'm
wrong). What I am proposing is that the title designator (T, R, et c)
should not be included in the .issues files unless the story has no title
in that issue. Damn, I hardly even understand what I am trying to say
myself here. What I would need to do is to come up with an actual example
from one of the files, but since I don't have access to many of those
books anyway, I'm not sure I can do that.
> My goal was to have all titles and descriptions in English, everywhere.
That's what I thought. But then we must ask ourselves: What is the
purpose of this database? If only to facilitate making references in
discussions on this mailing list, then this is certainly a good policy.
But when I downloaded the Swedish files, I did so in order to send them
to a comics collecting friend. Now, HE won't have any use for the English
titles. He will want the Swedish. Another reason that I am rambling on
about this subject is that I (perhaps foolishly) have undertaken the
mission of converting these files for the Grand Comics Database, which
has previously been discussed briefly on this list. For those who have
not heard of them, they are a bunch of lunat... err... devoted collectors
who are trying to index ALL American comics and making the results freely
available on the 'net. I thought that our database would fit in neatly in
that scheme if it can only be converted properly.
> Do you really think all story titles should be Swedish in the Swedish
> index?
YES! If I really want to discuss story D88215 (dummy example) I could
look the proper title up in the d.stories file. A little more trouble,
but not really that much, I think.
> > it's really not useful to know which language it originates from.
>
> Oh?
I just KNEW that statement would be misinterpreted. Let's take dummy
example D88215 for instance. Suppose John Anderson (isn't there some
Egmont writer by that name?) wrote the thing and called it "Day of the
Triplets". Then Stefan Dios translates it to Swedish and calls it
"Knattarnas stora dag". Finally, Anders Engwall indexes the Swedish
publication of the story and retranslates the title as "The big day for
Hewey, Dewey and Louie". Now, what is the point of knowing where this
last title originates from? Of course, we must point out that it is NOT
the original title, and this could be done by using a designator like Tr:
(meaning Translated). I think it's confusing enough having to keep track
of Sw, D and N (or whatever it is that you use). What of the day when
people will want to start indexing Russian, Italian, Chinese, Norwegian
and French books all at once?
> > ALL NON-ENGLISH issues should have ALL TITLES included in the issue files.
>
> Please propose a handsome lay-out for listing more than one title of a story.
&%$**! Without seeing my original context, I don't have a clue to what I
meant by that sentence, but I am pretty sure that I didn't mean what you
think.
Anders:
> And even *if* Harry's software could handle Latin-1, there's absolutely
> no guarantee that the nodes between you and him can.
Well, perhaps it was stupid of me to suggest Latin-1. There is absolutely
no realistic way to use it for plain-text files over the 'net. The only
way is to UUencode it. This could work, but it also means that all
indexers will have to know a lot about what they are doing. They must
know enough about Latin-1 to tell if that is what they are using (or else
a way of converting it) and also be able to UUencode. But I really don't
see another way to have Swedish, Danish and other non-English titles in
the database. Perhaps every language could use their own 7-bit standard,
but that would be messy for anyone interested in some other country's
titles...
Anyway, if that is any comfort, I could undertake to convert the Swedish
titles that are _currently_ in the database to Latin-1. But I would only
do it if we decide to go for Latin-1 on the whole line, since using it
for only one language makes no sense.
/F
PS. Thanks, Per, for telling me about DS. I had no idea it was
specifically Swedish. Guess I won't be using it any more in my
international messages. DS (oops!)
More information about the DCML
mailing list