CyberTaz said:
Hello, and thanks for taking the time to try, but all I see here is a
possibly plausible explanation of *why* the mail/browser devs did what they
did, but it in no way exonerates what they did. Further comments below;
OK, I need to try harder:
First, a little background:
Type and Creator is a Macintosh-only mess. In the days before OS X,
every file had two forks, Resource and Data. It was (generally) the job
of the Resource fork to hold metatdata - data about the data. The Type
and Creator codes therein were meant to be administered by Apple They
were used by the OS to choose an application to invoke on double-click.
Meanwhile, on the dark side, on unix and other operating systems like
VMS there was no resource fork to tell the operating system and its
built-in file system what application to activate to open the file. It
was never a problem in those far off times before the double-click and
the browser wars.
The user was supposed to choose the application from the command line.
You know, like run the application and then utter some kind of open
file command. e.g "MAR TESTPROG"
So, to deal with double-click on non-Mac systems, various fashions
developed over time. The common tricks were to infer type from the
filename (aka extension) and perhaps from a magic sequence of bits near
the front of the file's data.
Then OS X arrived, a unix at heart. Apple's file system had to work on
Classic files and still be usable by the unix utilities underneath.
What we have today is an unholy mess of hidden pretend resource fork
files and reliance on the other two methods. Type and Creator is pretty
much deprecated as a reliable concept.
So now we have a mail and web browser situation (on Macintosh only)
that needs a little ingenuity. An attachment arrives with only two of
its possible pieces of type evidence present. (Windows and unix and
everything else don't do hidden resource forks.) In the case of mailed
attachments, there may also be another piece of metadata, the MIME
type. So what's a poor browser or e-mail client or binary NNTP reader
to do? It is handed a blob of stuff from the world wide wibble and must
*create* a file that can be used by applications already on the machine
and by others that may arrive later.
Most Macintosh users do not choose their friends on the basis of what
OS they prefer, so 90-something percent of all attachments they receive
*contain no resource fork*. There is no type. There is no creator.
To keep some of those application authors from sleepless nights, the
mail, browser and NNTP applications *make* (not change) type and
creator codes from the available evidence - metadata in the filename
(aka extension) and magic strings in the preamble of the data, and
possibly a MIME type. They are NOT changing what is there. They are
guessing what could have been there if the file had originated on a
Macintosh. There is no reason for them to change what works. For the
ancient programs that need TYPE and CREATOR, the old 'uns are the good
'uns. New applications know about the mess, and know that the TYPE and
CREATOR may be unreliable once the file has passed through a non-Mac
system.
If MS were the only one to have ever changed a file type this might be a
significant point, but that's hardly the case
IOW they took the liberty of modifying the files they were entrusted to
deliver intact in order to cover their own butts? They might just as well be
entitled to change the language of the content because the recipient "may
not speak the language" used by the originator. I'm sorry, perhaps I'm a bit
na•ve, but when I entrust a carrier to deliver my goods I expect them to
deliver what I sent - not "adjust" it to suit their own special interests.
If I order bacon with my eggs I don't expect the server to substitute ham
because s/he's afraid I may have a cholesterol problem. In this case, if I
ship a W8BN that's what I expect to have delivered. If the recipient can't
work with it that's between "me & he". For the handler to *arbitrarily*
change that is just plain wrong - and in this case the liberty taken
backfired on them.
As the backgrounder above shows. They had no choice. There was no type
and creator in 90% of files received over the net. Perhaps there should
be a ???? ???? Oh wait! There is! That was guaranteed to have *no*
preferred application on OS 9 and before.
Again, that begs the question as originally posed & circumnavigates the case
in point: The "backdating" should never have been a concern because it shold
never have been done.
I don't think so. It *had* to be done when no type and creator was
present every time the file had the misfortune to pass through a
non-Mac system.
Well, it wouldn't have been caught in the Beta because the change wasn't
implemented until the 12.1.0 update
But I still say it isn't the
responsibility of *any* software developer to test for whether other
software is going to screw with their files - at least not as a preliminary
priority.
Exactly. The product was rushed out AND the update was rushed out.
I'm sorry. It is Microsoft screwing with their own files. They *knew*
that net-transferred content was bound to contain type and creator of
doubtful provenance because *their* operating system was the prime
culprit of not putting it in or of ripping it out. It was a silly
mistake to tighten up something that was already broken, and had never
been used seriously except on Mac systems more than 6 years old.
That we don't - and may never - know, so to dismiss it as "pathetic" and
"minor" is just as inappropriate as declaring it a matter of national
security... But that's a moot point. It's well within a developer's purview
to support or not support any file types they deem appropriate. In fact,
it's rather ironic to argue that they don't in support of other developers
having the right to make the decision for them.
Bob, you could call it GWOT (Global War On Terr^h^h^h ype codes)
OS X *is* the operating system without hundreds of thousand of viruses
in the wild. That qualifies it as minor, when taken in conjunction
with..
...."pathetic", because the user can immediately open the file directly
from Word. In other words, if there were a real threat, the response
was utterly inadequate.
Just for the sake of playing devil's advocate: I think you may give far too
much credit to the creators of Firefox, et al. It can just as easily be
argued that they chose WDBN due to error or ignorance & have failed to
correct it due to ignorance or neglect. Remember that Mozilla 1.0 was
released in June of 2002, at which time WDBN had been abandoned for nearly 6
years... And don't the creators of software that handles our files have some
responsibility to make sure their products don't interfere?
OK, you probably see why, now I have shown you the history of it. It
was neither ignorance nor error. It was reasonable to assume that
Microsoft would adapt their file opening security processing to deal
with the dodgy provenance of type and creator on files that had passed
through the internet. It is reasonable of Firefox et. al. to not break
existing *old* software. It was reasonable of them to make a best guess
of what type and creator would best suit applications that were written
BEFORE internet use became common, and to let newer applications behave
as though they had heard of the internet and the general mess-up of
type and creator.
This whole debacle reminds me of the story about the woman who insisted that
her neighbor be arrested for indecent exposure. She explained to the police
that "All you have to do is climb up on the fence, lean over against his
house, stick this screwdriver in the gap on the edge of his window & pry
back the slats of his blinds while he's undressing & you can see everything
he has."
Once again, you analogy is poor. If it were the neighbour who phoned
the police to complain of his own behaviour, just in case his neighbour
pried open his slats.....
Perhaps this is one of those issues on which we'll have to agree to
disagree
At one point I thought that the change in behaviour was MS covering
their six, but the more I think about it, the only possible explanation
is they drew their SP1 revolver and shot themselves in the foot.
But they were awfully quick on the draw. After waiting 12 years for
everyone to lose interest in the gunfight.
Remember. When a double-clicked file is handed to Word, it has all the
evidence about its provenance that the mail client or browser had. Why
oh why does it suddenly throw a hissy fit if it encounters one of its
own ancient type codes? A few seconds later it will behave properly
when the file is opened explicitly, whether or not the file is well
formed.
