Andreas J Guelzow <> wrote on 03/15/2009
11:31:34 PM:
>
> On Sun, 2009-03-15 at 22:39 -0400, wrote:
> > To accomplish there we need to either ensure that producers write the
data
> > consistently, or that consumers use a more complicated heuristic to
> > determine document type. I'm inclined to believe that this will work
best
> > if we require both.
>
> I find it greatly disturbing that the standard would try to force an
> application to show certain latin letters in a file name just to have
> that file to be considered a conformant document.
>
That's the way the dominant operating system on the planet does it. I
find it disturbing as well, but it is the least-bad approach that works
with all of the OS's is broad use today. If you have a better idea, then
I'm all ears. But requiring an operating system to unzip a file and parse
the XML in order to determine whether to show a text icon versus a
spreadsheet icon for the file in a file explorer GUI -- that isn't going
to work well.
> Perhaps if you weren't used to writing in the latin alphabet it would be
> more obvious that having such cryptic (and in the user's language likely
> meaningless) letter combinations attached to a file name.
>
End users don't necessarily need to see or deal with file extensions. For
example, Windows, since XP at least, hides extensions by default in the
GUI. It is still a hack and not as clean as other OS's. But I don't see
how to avoid dealing with it in ODF.
And remember, the file extensions are already part of our MIME content
type registration. So you'll see Apache servers set up to serve up ODF
documents with the correct MIME type based on looking up the file
extensions. That breaks if documents are not saved with consistent file
extensions in the first place. It is all one giant hack, but a
well-understood one at least. I'm not sure there is much we can do about
it.
Sure, we could eliminate this from the conformance requirements. But if
someone takes a ODF presentation file and renames it with an .ods
extension, it will display with the wrong icon in Windows, it will be
served up with wrong content type in Apache, it will launch with the wrong
application if you receive it in an email attachment, and it will do
strange things if you try to load it in Gnumeric or any other
ODF-supporting spreadsheet. Personally I think this kind of inconsistency
is something we should try to discourage, and a conformance requirement in
favor of consistency is the way to do it.
-Rob
