Re: Proposed issue: site metadata hook

On Thursday, February 13, 2003, 9:35:03 AM, Patrick wrote:




>> -----Original Message-----
>> From: ext Chris Lilley [mailto:chris@w3.org]
>> Sent: 13 February, 2003 01:40
>> To: Stickler Patrick (NMP/Tampere)
>> Cc: timbl@w3.org; www-tag@w3.org
>> Subject: Re: Proposed issue: site metadata hook
>> 
>> 
>> On Wednesday, February 12, 2003, 10:52:09 AM, Patrick wrote:
>> 
>> >> The solution should also allow a distributed and scalable 
>> architecture
>> >> that allows multiple users to share one server, rather 
>> than the wildly
>> >> unrealistic and unfair constraint of the server owner controlling
>> >> everything.
>> 
>> PSnc> While I'm all for the most flexible and enabling 
>> solution, I think
>> PSnc> this issue of some users not having control over their 
>> servers is
>> PSnc> a social issue, not a technical issue.
>> 
>> That is where you would be wrong. Fixing it (by for example giving
>> everyone a free server they control) would be a social issue. Failing
>> to account for this restriction in a technical architecture is a
>> technical issue.

PSnc> But *removing* such restrictions (which I presume you mean by
PSnc> "taking account" of them) is a socially motivated act.

That assumes that the restrictions were placed there as a deliberate
act rather than by hapenstance.

PSnc> What you are talking about is using technical means to affect 
PSnc> a social situation by having the "solution" impinge upon the
PSnc> rights of the server *owner*.

No, I am not. But you are.

PSnc> The technology should not mandate such social issues.

I agree, so your proposal to enforce that anyone who does not own a
server cannot have any metadata is an attempt to impose a social issue
and should be resisted.

PSnc> Since any users is *free* to get their very own server, then
PSnc> defining the technology in terms of the rights of the server
PSnc> owner is neither a social nor a technical problem.

>> PSnc> Folks that live in apartment buildings don't have total control
>> PSnc> over their domiciles. If they want total control, they should
>> PSnc> buy a house.
>> 
>> I am not especially interested in building a web architecture that
>> only applies to corporate brochure publishers and ignores anyone else
>> as irrelevant.

PSnc> Sheesh. You must be kidding!

No, actually.

PSnc> Exactly *how* does the present web architecture *ignore* the 
PSnc> "common man", eh comrade?

Since you resort to denigration and ad hominem I assume that means
you have run out of technical arguments.

PSnc> There is no dark sinister society preventing every Joe on the
PSnc> street from having their own server. Criminy, just get a
PSnc> fixed line connection and put Apache on your home computer!

PSnc> Oh, you say it's too expensive? Well..... So are ten bedroom
PSnc> homes on twenty acre lots on the ocean front!

PSnc> Welcome to the real world.

Thanks for sharing that wonderful vision with us.

PSnc> If there is some technological aspect that makes it cost
PSnc> prohibitive to *own* a web server, then certainly, let's
PSnc> address that.

No, lets just adress that the entirety of the space on a web site
might not have the same metadata or the same policies. Take individual
users out of the equation and substitute different departments of one
corporation, if it makes you feel more comfortable not having to deal
with the general public.


PSnc>  But the web architecture should not in any
PSnc> way impinge upon the rights of the *owner* of that web
PSnc> server, and the owner is free to restrict "tenants" of
PSnc> that server however is agreed to within the domain of
PSnc> that server.

>> PSnc> And if enough tenants in a given apartment building 
>> (enough users
>> PSnc> of a given server) want some change to the shared 
>> resources, then
>> PSnc> they can request it, or demand it, or then find a more suitable
>> PSnc> place to live.
>> 
>> Thats a social issue ;-)

PSnc> Duh. And one that the web *architecture* need not "address".

>> PSnc> Whatever solution is promoted, it should not disregard 
>> the rights
>> PSnc> of server owners to decide how their servers are used,
>> 
>> Sure, server owners should be able to lock down a server if they want
>> to. But that should not be the default option and not the only option.

PSnc> Fair enough. But you seem to be implying that the web architecture
PSnc> should allow a user to override the rights of the web site owner.

No, *you* seem to be assuming that and getting hot under the collar
about it.  What I am assuming is that the right of the 'portion of a
website' owner might not, in all cases, be autocratically overridden
by the owner of the entire server. And in the cases where it is not,
how to say what the policies of the subpart are? Clearly, requiring
all the tenants to append onto one master file in the root directory
is a poor solution I think you would agree.

After all the "owner" of the server that you seem so keen to protect
has already allowed their 'tenants' to provide content in their
directories. Allowing them to also provide metadata about (and
possibly in) those directories is hardly a bolshevik revolution.

PSnc> That
PSnc> if the web owner says "no robots here" that the user should still
PSnc> be allowed to invite them in. That's what you suggested in your
PSnc> original post.

PSnc> That's like sneaking you friends into the theatre by the back door.

PSnc> Shame on you. (you were a mischievous kid, weren't you ;-)

Once more the stimulating riposte.

>> PSnc> and the
>> PSnc> fact that folks residing on those servers have agreed to waive
>> PSnc> certain rights in order to reside there.
>> 
>> No, they have not so agreed. 

PSnc> Eh? Sorry. Surely you don't expect me to dig out and provide
PSnc> you with examples of the legal agreements typical of most
PSnc> web hosting services.

No, I just expect you to use logic. You are in effect saying that
no-one can put private information on their website and control
whether it gets crawled.

PSnc> Really, Chris, I think you're view on this is very distorted
PSnc> and unrealistic.

Thank you for your opinion.

PSnc> Users most certainly *have* agreed to such restrictions.

Look, you may choose to characterise me as a mischeivous child, but I
was *there* when the robots.txt thing was being worked out and believe
me, it was a quick and dirty hack that just caught on. Attempting to
imply

>> And an unrealistic, elitist and utopian
>> architecture that applies only to some small fraction of content
>> providers is just a waste of time.

PSnc> Onward the revolution...  ;-)

Pragmatism not politics. A web architecture that only applies to 10%
of the web is a pointless waste of time.

>> Besides, the same problem afflicts corporates too so its not some
>> bleeding heart liberal philosophy here - its pragmatism.

PSnc> I'm all for an architecture

Ah, okay, lets leave mention of actual people out of this and just
pretend its corporate departments. That seems to make you both happier
and more technical in your responses.

PSnc> that allows for site owners to
PSnc> empower their tenants in whatever way *they* see fit -- including
PSnc> the ability to selectively specify subsets of the site visible
PSnc> to robots, crawlers, etc. and leaving such visibility up to
PSnc> the discretion of each tenant, etc., but only *if* the web site 
PSnc> owner decides to do so.

And if they do, are they likely to allow all those tenants to all go
editng one master file when most of that file does not concern them?
Now who is being unrealistic?

PSnc> But I am definitely opposed to any architecture that allows a
PSnc> tenants rights to override or circumvent those of the site owner.

PSnc> I do very much hope you would agree with at least that last
PSnc> point.

Partially.

Lets consider an architecture where the corporation owns / and accounting
owns /corporate/accounting and marketing owns /comm/pr

Lets assume that the corporation decides that it does not want /
crawled. Lets assume that marketing wants /comm/pr crawled.

You seem to worry that I want to empower marketing to override the
settings on / wheras in fact I want them to be able to control their
own little bit and not be able to control everyone elses bits.

PSnc> And as a final point, wouldn't the ability to express all the
PSnc> complexity of site configuration and the rights of tenants,
PSnc> etc. be so much easier if one could just use RDF,

I don't recall excluding this.

PSnc> and then
PSnc> just ask the site to tell us if e.g. a robot can inspect
PSnc> the web space of tenant "John Doe"...???

And you would do that how?

PSnc> Why do we need anything more than the semantic web extensions
PSnc> to the present web architecture

If I knew clearly what those were then I might be able to answer you.
But at present there does not seem to be a list of them.


-- 
 Chris                            mailto:chris@w3.org

Received on Thursday, 13 February 2003 08:02:37 UTC