OT: GUID's

  • Thread starter Thread starter Bob
  • Start date Start date
B

Bob

We have Verisign which serves the internet domain database. How about a
service that delivers something more general and powerful, like metadata on
GUID's for... well, everything? It could not only help distribute commonly
used files, but provide a stronger organizing force for the web and give
search engines a common, more coherent foundation. GUID registration fees
could cover the service and maintenance of the metadata. Just a thought.

Bob
 
Think of it as a Global Assembly Cache where "Global" really IS global. It
could serve metadata about assemblies and where to locate them on an
internet-wide scale. And why stop there? How about fonts. And domain names.
And all electronically available publications. And businesses. With the use
of a bit of library, business, internet and crytographic science I think we
could make something very interesting: a decentralized, authoritative
database on everything.

Bob

Well Bob...

Maybe a "for instance" might help :-)
 
Ok... so you've assigned a GUID to http://www.cnn.com current page? Did it
include the dynamic ad on display at the moment you assigned the GUI, the
textual content or some other combination? Tomorrow when the page is
different do they submit a request for a new GUID or just update their
metadata information?

So when I press "send" is this message assigned a GUID automatically or do I
fill something out first explaining the contents so that searching is
easier?

Does the metadata on "crop circles" and "communicating with the dead" list
it under "science" or "scam" and who decides? If it is the guy
communicating with the dead how is this considered "authoritative"?

If we can obtain software, (and fonts) and articles and such currently what
is it your GUID gets me other than a 16-byte identifier of course? When I
tell somebody "buy WinZip" are you thinking I might suggest the purchase of
F9168C5E-CEB2-4faa-B6BF-329BF39FB1E5 instead? What if that version has been
replaced or the link that your database contains is simply out of date?

I think your idea of cataloging everything is interesting ... or should I
say is F9168C5E-CEB2-4faa-B6BF-329BF39GA1D6 and when you click on that the
definition of "interesting" shows up in your browser... :-) and D7 would
have it in Spanish, and D8 would have it in French, etc., etc.

Bob... what are you talking about?
 
Simplest example: say I did want to look for "WinZIP" and download a trial
version. I can find the official website and download it. But what about
more obscure files? What if there is no official website? I may find the
obscure file with a search engine, but how do I check to make sure I was
actually getting the real thing, or the right version? Sure, if it's a
DotNet compile I can see, but how about a mechanism that is independent of
DotNet, applicable to any file, that requires no alteration? Enter a central
authority, much like ICANN, coupled with exiting search engines. Files would
be registered with a unique identifier and a hash code. With a such a
service, I could just type the file name and sort the search hit results by
registered file creation date, number of sites where it's available, ID of
the original registrant, etc, and have automatic validation on download. You
could have any software with an auto-update feature just like Microsoft's,
without any need for any particular dedicated website or file format. I'm
sure you can think of a host of practical issues, but I don't think the idea
is completely unworkable.

Now, more of a stretch. What if we forget about files and talk about
concepts. Think about every concept in the book, from crop circles to
pottery. Everything, independently of language. Give everything you can
think of a GUID. Forget about how you associate these concepts to one
another, categorize them usefully, or manage them for a user. It's huge, can
be done, has been done, but it's not important. A web site administrator
could search through this GUID database, find the concept he was looking for
(in Russian, English, whatever), and select the GUID's he thinks his website
has relevance to, and post these GUID's in a manifest, maybe in web page
headers, maybe as direct markups to sections of text or pictures or any
other file (see how this is a superset of the previous example?). Sure,
you'd get abuses and bad links just like now. But imagine the improvement in
accuracy for all the good search hits if this became widely adopted. Imagine
the birth of a new language-independent Framework - not in code, but in
human knowledge.

Bob

Ok... so you've assigned a GUID to http://www.cnn.com current page? Did it
include the dynamic ad on display at the moment you assigned the GUI, the
textual content or some other combination? Tomorrow when the page is
different do they submit a request for a new GUID or just update their
metadata information?

So when I press "send" is this message assigned a GUID automatically or do I
fill something out first explaining the contents so that searching is
easier?

Does the metadata on "crop circles" and "communicating with the dead" list
it under "science" or "scam" and who decides? If it is the guy
communicating with the dead how is this considered "authoritative"?

If we can obtain software, (and fonts) and articles and such currently what
is it your GUID gets me other than a 16-byte identifier of course? When I
tell somebody "buy WinZip" are you thinking I might suggest the purchase of
F9168C5E-CEB2-4faa-B6BF-329BF39FB1E5 instead? What if that version has been
replaced or the link that your database contains is simply out of date?

I think your idea of cataloging everything is interesting ... or should I
say is F9168C5E-CEB2-4faa-B6BF-329BF39GA1D6 and when you click on that the
definition of "interesting" shows up in your browser... :-) and D7 would
have it in Spanish, and D8 would have it in French, etc., etc.

Bob... what are you talking about?
 
That's good... an example helps where idle chat doesn't. So playing devil's
advocate since you aren't actually outlining how anything works :-)

Bob said:
Simplest example: say I did want to look for "WinZIP" and download a trial
version. I can find the official website and download it. But what about
more obscure files? What if there is no official website? I may find the
obscure file with a search engine, but how do I check to make sure I was
actually getting the real thing, or the right version?
Click to expand...

Most non-WinZip apps would be whatever the author decided was the "official"
website. What is the definition of "the real thing"? You mean "the one
that your GUID has been issued to" what are the criteria for you to issue
such a GUID? Are you proposing testing all the software first? You
wouldn't issue a GUID to software which contains a virus set to go off in 30
days? They mentioned this when the request for the GUID was made?

Is it an authentication system that you are proposing? So somebody named
"Bogus" submits a request for the "CoolDude" app for Linux... take us on a
tour of the process you are going through before you issue the GUID.
Files would be registered with a unique identifier and a hash code. With a such a
service, I could just type the file name and sort the search hit results by
registered file creation date, number of sites where it's available, ID of
the original registrant, etc, and have automatic validation on download.
Click to expand...

Again I have to guess you have more figured out than you've explained. How
did all the sites get into your database? Why can't I simply provide a link
to my utility or for that matter WinZip without telling you? How do you
resolve broken links any better than Google? Do companies no longer
offering the CoolDude download feel an obligation to let you know?
Automatic validation... why not just credit my account for frequent flier
miles while you are at it? Can you explain how your system controls the
validation? I click "Save Target As..." and?
You could have any software with an auto-update feature just like Microsoft's,
without any need for any particular dedicated website or file format. I'm
sure you can think of a host of practical issues, but I don't think the idea
is completely unworkable.
Click to expand...

Exactly... yet Microsoft doesn't subscribe to the universal GUID system. I
don't doubt that people could download software, I am pointing out that they
already can without a central authority.
Now, more of a stretch. What if we forget about files and talk about
concepts. Think about every concept in the book, from crop circles to
pottery. Everything, independently of language. Give everything you can
think of a GUID. Forget about how you associate these concepts to one
another, categorize them usefully, or manage them for a user. It's huge, can
be done, has been done, but it's not important. A web site administrator
could search through this GUID database, find the concept he was looking for
(in Russian, English, whatever), and select the GUID's he thinks his website
has relevance to, and post these GUID's in a manifest, maybe in web page
headers, maybe as direct markups to sections of text or pictures or any
other file (see how this is a superset of the previous example?).
Click to expand...

Bob... what is "everything" when you talk about assigning a GUID? And why a
GUID... they aren't something invented last year. It could have done a
century ago, you think they didn't have 128-bit values? Where do you think
the name Google comes from?

So... the administrator searches for "VB.Net" and he gets a list of 30
million GUIDS? Does that sound even remotely useful? Do you think
F9168C5E-CEB2-4faa-B6BF-329BF39GA1D6 will be useful to you? You can't know
of course but tell me if it "sounds" useful? Does it sound more useful than
say F4168C5E-CEB2-4faa-B6BF-329BF39CB2E7?
But imagine the improvement in
accuracy for all the good search hits if this became widely adopted. Imagine
the birth of a new language-independent Framework - not in code, but in
human knowledge.
Click to expand...

When you put it that way, I see it all now. It was never that there was bad
information floating about or no way to catalog it we just didn't have
everything assigned a GUID before. How do you imagine everybody overlooked
such an obvious solution?

Bob: Seriously I'm about the only one entertaining your idea so bear with
me on this. What are you talking about? :-) How about everybody who sells
goods and services posts their price in your universal GUID database and
that way anybody can simply cost compare across the entire world? Like all
the manufacturers and stores and even E-Bay will post and update your
database... that would be cool right? Why on Earth wouldn't they want to do
this... you're uh, Bob after all :-)

Doesn't it sound remotely impossible? What do you suppose your monthly
costs for server hardware and maintenance personnel will be? Where do you
derive the income?

Tom
 
Most non-WinZip apps would be whatever the author decided was the
"official"
website. What is the definition of "the real thing"? You mean "the one
that your GUID has been issued to" what are the criteria for you to issue
such a GUID? Are you proposing testing all the software first? You
wouldn't issue a GUID to software which contains a virus set to go off in 30
days? They mentioned this when the request for the GUID was made?
Click to expand...

There would be no validation, just a control in place to prevent
overgeneration: a small fee, or some other mechanism.
Is it an authentication system that you are proposing? So somebody named
"Bogus" submits a request for the "CoolDude" app for Linux... take us on a
tour of the process you are going through before you issue the GUID.
Again I have to guess you have more figured out than you've explained. How
did all the sites get into your database? Why can't I simply provide a link
to my utility or for that matter WinZip without telling you? How do you
resolve broken links any better than Google? Do companies no longer
offering the CoolDude download feel an obligation to let you know?
Automatic validation... why not just credit my account for frequent flier
miles while you are at it? Can you explain how your system controls the
validation? I click "Save Target As..." and?
Click to expand...

You seem to miss the point that file validation is one of the major reasons
for doing this. OK, we'll go with CoolDude. Say he registers a unique ID for
his product with this hypothetical central authority for $1.00, puts it out
on the web. Other people like it and share it on their websites. Then a year
later the guy packs up his website and leaves for... uh... Botswana... never
to be heard from again. Meanwhile, web crawlers have been aware of the
unique identifier database. They've been crawling the web and cataloguing
the markup GUID tags. Later, Joe User does a special search of the central
authority database through his favorite search engine, recognizes the
description and signature of what he's looking for, and then clicks on a
link to download the file. The search engine could then poll the different
sources for file size, eliminate bad sizes or changed dates, download it to
his computer, then verify that the file matches its authoritative hashcode.
Exactly... yet Microsoft doesn't subscribe to the universal GUID system. I
don't doubt that people could download software, I am pointing out that they
already can without a central authority.
Click to expand...
Bob... what is "everything" when you talk about assigning a GUID? And why a
GUID... they aren't something invented last year. It could have done a
century ago, you think they didn't have 128-bit values? Where do you think
the name Google comes from?
Click to expand...

This could not have been done a century ago. Handling and serving a massive,
language-independant database requires computers, and in the context of this
discussion, the internet.
So... the administrator searches for "VB.Net" and he gets a list of 30
million GUIDS? Does that sound even remotely useful? Do you think
F9168C5E-CEB2-4faa-B6BF-329BF39GA1D6 will be useful to you? You can't know
of course but tell me if it "sounds" useful? Does it sound more useful than
say F4168C5E-CEB2-4faa-B6BF-329BF39CB2E7?
Click to expand...

That's the point of coupling a GUID database with metadata and a search
engine. You seem to be stuck on people typing in GUID's. A useful system
would almost never show end users GUID's unless they wanted to see them.
When you put it that way, I see it all now. It was never that there was bad
information floating about or no way to catalog it we just didn't have
everything assigned a GUID before. How do you imagine everybody overlooked
such an obvious solution?
Click to expand...
Bob: Seriously I'm about the only one entertaining your idea so bear with
me on this. What are you talking about? :-) How about everybody who sells
goods and services posts their price in your universal GUID database and
that way anybody can simply cost compare across the entire world? Like all
the manufacturers and stores and even E-Bay will post and update your
database... that would be cool right? Why on Earth wouldn't they want to do
this... you're uh, Bob after all :-)
Click to expand...

OK, we're talking about the more general case now, and not just files.
Prices and everything else site-specific would not be stored in a
language-independent, "concept" GUID database. Again, the web crawlers would
simply detect found markups in web pages, stupidly, and preserving old and
bad links just like they do now. But at least the search hits based on these
markups would no longer depend on words and break on misspellings or
alternate phrasings. And I would be able to tell that a certain page I
found, in Spanish, was about cookery, *before* I bothered to translate it or
learn any of the relevant Spanish words and phrases. Yeah, there would be
bad matches, but that wouldn't necessarily make the whole sytem useless.
Doesn't it sound remotely impossible? What do you suppose your monthly
costs for server hardware and maintenance personnel will be? Where do you
derive the income?
Click to expand...

I find it hard to believe that you can't look at how existing search engines
operate to get an idea.
 
Back
Top