Re: Hash vs Hashless URIs

On 11/21/12 8:35 AM, Andrei Sambra wrote:
> On Wed, Nov 21, 2012 at 7:30 AM, Kingsley Idehen 
> <kidehen@openlinksw.com <mailto:kidehen@openlinksw.com>> wrote:
>
>     On 11/21/12 12:24 AM, Andrei SAMBRA wrote:
>>     On Tue, Nov 20, 2012 at 9:45 PM, Kingsley Idehen
>>     <kidehen@openlinksw.com <mailto:kidehen@openlinksw.com>> wrote:
>>
>>         On 11/20/12 6:22 PM, Melvin Carvalho wrote:
>>
>>             OK, sorry for being a bit slow.  Proxy URI's do make a
>>             lot of sense.  I've read most of your posts but if you
>>             had a pointer to refresh my memory that'd be great.
>>
>>
>>         When I say proxyURI I also mean a wrapper URI. An
>>         intermediary URI that handles all the data access and
>>         transformation heuristics via URL patterns. For example, how
>>         we (and others) use existing data formats and protocols to
>>         produce 5-star Linked Data [1][2][3][4][5] on the fly. Others
>>         have produced similar and a majority are based on hashless URIs.
>>
>>         Instead or re-beating this matter to death, I really
>>         encourage the group to move to a proper vote on this matter.
>>         I don't feel we are actually making any progress on the
>>         debate front. It's time to vote on this matter so that we can
>>         all make clear decisions about our respective strategies and
>>         priorities, bearing in mind general time scarcity.
>>
>>
>>     Kingsley, I'm starting to get the impression that you are going
>>     around in circles. One day you agree about # URIs, one day you
>>     don't.
>
>     When did I ever give you the opinion that I support constraining
>     WebIDs to hash based HTTP URIs? Can you point to any resource
>     where I've indicated what you claim?
>
>
> It might have been just an impression, I admit. I'll take it back 
> since it was not the case.

Look, I have hash and hashless URIs. I've never designed anything that 
prefers one over the other. The same applies to all the RDF model 
serialization formats and markup syntaxes. All examples of my CEO 
myopia, clearly.

My focus, interest, and passion are always centered on genuine openness 
via industry standards. That's it.

>>     Your concern, which is also mostly shared by people affiliated to
>>     OpenLink Software, deals with compatibility with your current
>>     implementation/products.
>
>     You really think it boils down to that, seriously? I hope you
>     understand that Linked Data isn't some new fad. Neither is the
>     WebID concept. Please invest sometime in actually looking the
>     history of both. Google is your friend on that front, at the very
>     least.
>
>
> Yes I do think that. Why would looking up Linked Data and WebID be 
> related to this topic?

Because a WebID has to be a 5-start Linked Data URI. As stated above, 
our products produce Linked Data, they don't give a darn about URI 
schemes, HTTP URI styles, or RDF serialization and syntaxes.

You seem to subscribe to the broken doctrine: if you aren't with us 
(hash HTTP URI supporters), you must be against us. The doctrine is 
eternally broken. Antithetical to essence of the AWWW.


>>     While it _is_ a valid reason, worthy of being taken into
>>     consideration, I personally believe that it is not sufficient to
>>     dismiss the current proposed spec, which has the potential to
>>     affect many more future users and applications.
>
>     Raising concerns isn't dismissal. If anything you and the others
>     who are supporting a flawed view are the one's who are in fact
>     being dismissive. Again, do understand that TimBL explained his
>     concerns about hashless URIs, what he didn't do is imply that
>     WebID MUST be scoped to said HTTP URI type. There's a different
>     between implementation optimization and architectural design.
>
>
>  The issue here is not about being dismissive. By saying that it is 
> clear that you are taking this personally, which is not the case.

You are making personal statements.
You don't seem to be able to distinguish fundamental debate and personal 
opinion.

>>
>>     With the risk of repeating myself yet again, I want to say that
>>     the goal of the WebID CG is to move forward and become a WG. We
>>     have been advised by a lot of experienced people at TPAC that the
>>     best way to proceed is to eliminate all unnecessary ambiguity
>>     from the current spec. You need to understand that in this case,
>>     adding a few constraints helps us to achieve this goal. This
>>     matter is not about personal preferences, but about valid reasons
>>     why several decisions were taken, decisions which are above all
>>     about scalability/performance.
>
>     Since you clearly have no idea why I am raising these concerns,
>     and you've opted to take the cheap personal accusations route, be
>     rest assured you are burning up my patience. I do have much better
>     things to do with my time.
>
>
> Please forgive me, I was not trying to accuse anyone at this point.

You made a blatant accusation.

> I was just explaining why we're pushing for these constraints. You are 
> taking it personally.

You made a personal accusation instead of making a counter point. As I 
said, there's no shortage of history and literature re. these matters.

>>
>>     My impression of you at this point is that you're viewing the
>>     whole process of having WebID reach a TR through the eyes of a
>>     CEO, very concerned about one of your products.
>
>     See my last comment above.
>
>
> Your main argument so far was that you have a 30k user base and you 
> need to make sure they can still use your products. Read the minutes 
> again.

I said: we already have 30K+ WebIDs that have been issued. I didn't say 
that we have 30K worth of hashless URIs issued. My comments have these 
not so obvious dimensions:

1. implementer/maker weight -- if OpenLink Software has implemented 
WebIDs that are actually used by our customers, it doesn't count for 
zero (we've are not experimenting with WebID)

2. actual interop -- there are some of the WebIDs that have been issued 
that are hashless, and that isn't unique to us .


>
>>     The standardization process is a long a tedious one.
>
>     See my last comment above, and invest a modicum of time understand
>     who I am and where I come from. You are beginning sound quite
>     immature and utterly inexperienced re. W3C process and technology
>     architecture!
>
>
> That is your personal opinion. You are never referring to the spec 
> whenever you disagree with something.

What spec(s)?

> Have you thoroughly read the spec lately? The feedback I got so far 
> was that it is very easy to understand and that it makes a lot of sense.

Being easy to understand isn't the goal. Being useful, implementable, 
and architecturally defensible is the goal.

> It's important we stop talking out of context and we refer to the spec.

Like I said, time to call a vote on this matter. I have much better 
things to do with my time. WebID has far less effect of me and OpenLink 
Software that you'll ever really understand, so maybe you will find that 
out a different way, in due course.


All:

Please call vote on this matter. I have better things to do with my 
time. If a majority want a WebID to be defined a hash based HTTP URI, 
let's get that in the open right now. We all have decisions to make 
about how we spend our time and priorities across our respective 
development efforts.


Kingsley
>
> Andrei
>
>
>>     To be able to succeed, you need to realize that by making a
>>     compromise at this point (by adding constraints to the spec) we
>>     can finally move forward. Please, try to see the bigger picture.
>
>     See my comments above.
>
>     -- 
>
>     Regards,
>
>     Kingsley Idehen 
>     Founder & CEO
>     OpenLink Software
>     Company Web:http://www.openlinksw.com
>     Personal Weblog: http://www.openlinksw.com/blog/~kidehen
>     <http://www.openlinksw.com/blog/%7Ekidehen> Twitter/Identi.ca
>     handle: @kidehen Google+ Profile:
>     https://plus.google.com/112399767740508618350/about LinkedIn
>     Profile: http://www.linkedin.com/in/kidehen
>
>


-- 

Regards,

Kingsley Idehen 
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen