Re: What is a failure?

On 2014-12-10 20:38, Michael Champion (MS OPEN TECH) wrote:
>> For what reason are the technologies listed there called "failures"?
> https://www.w3.org/wiki/AB/2014-2015_Priorities/w3c_work_success is a wiki, feel free to edit it.  The initial contents were off the top of my head, and were somewhat deliberately controversial in order to get some discussion started.   As I recall my thinking, both WS* and the Semantic Web have been successful in terms of generating W3C standards, but neither lived up to the hype with which they were launched some years ago.

The wiki ... until I get a better feel for how people approach this 
area, I stay on the mail ;-)

One could see an analogy between on the one hand the way certain 
technology comes to novel uses, and on the other hand, the concept 
"exaptation" in biological evolution. For the latter, the Wikipedia 
article starts as:

*Exaptation* (a replacement for the teleologically 
<http://en.wikipedia.org/wiki/Teleology>-loaded term 
"*pre-adaptation*"^[1] 
<http://en.wikipedia.org/wiki/Exaptation#cite_note-Gould1982-1> )
        and the related term *co-option 
<http://en.wikipedia.org/wiki/Co-option_%28biology%29>* describe a shift 
in the function of a trait during evolution 
<http://en.wikipedia.org/wiki/Evolution>.
        For example, a trait can evolve because it served one particular 
function, but subsequently
        it may come to serve another.

We do see things like this in many technological areas, where in effect 
good use of the technology is made in other domains/use cases than 
originally envisioned. Well, Flickr originally started as a platform for 
multiplayer online games, with support for uploading images to embellish 
your persona. Soon it was noticed that people were using the platform to 
share images, and based on that, the path of Flickr was redirected to 
photo sharing service.

The semantic web has resulted in much technology that is used by people 
who are not interested in the philosophical dimension of semantics and 
ontology, but rather see the technology as a generic multi-purpose 
technology that provides extremely good support for information exchange 
in distributed / decentralised applications (e.g., using SPARQL).

So, yes, the Semantic Web has not materialised as originally envisioned 
("intelligent exchange and inference of knowledge/information"), but it 
has become a platform for loosely coupled systems.

In addition, in many areas the information representation format (RDF) 
has replaced the structured databases (e.g. relational databases), 
especially in contexts where information handled is not always in a 
fixed structure (like tables).

So, "success" and "failure" is actually often "in the eyes of the 
beholder". Which means that it is often important to be explicit about 
the point of view from which one judges X as a success or as a failure.

>
> I do like the distinction between:
>> (1) failure to reach the formal objectives specified for the work.
>> (2) failure to achieve the expected real-world effects
> I personally think this task force should prioritize (1), because if an effort can't even get enough support to advance through W3C, it's not likely to be destined for real world success.   Or something might have so much real-world success as a de facto standard that there isn't much interest in formalizing W3C Recommendation.  In either case, it's probably not a good use of W3C's resources to keep investing in that area.
>
> That's not to say we should ignore (2), just be fairly humble about the prospects of predicting real world success/failure given all the extraneous factors that go into it, e.g. the whims of what the technorati consider in fashion this season, and random shocks such as the Snowden revelations that drove a lot of rethinking about security/privacy.

Agreed. In general, only time will tell whether a certain W3C technology 
will become a critical component of the web.

Evaluation before starting work and during work is probably the only 
thing that makes sense. And for that we need some conceptual framework, 
based on which one can define rules and guidelines for how to evaluate 
proposed or on-going work.

And, of course, no such evaluation method is fool-proof. Was it not the 
case that XHTML looked good when proposed and started, as well as during 
its first phases of life. In effect, the torpedo that sank XHTML came 
from the outside (WHATWG). So either one has to acknowledge that certain 
failures cannot be anticipated ("shit happens, we have to live with 
it"), or one has to spend a lot of effort to monitor and evaluate what 
is going on in the external world, and proactively take action 
(stop/redirect work) instead of reactively take actions.

Perhaps one could say that if W3C *never* experiences failures, then 
W3C  was not ambitious enough.

What are the practical differences between evaluating proposed work 
("Should this proposed WG on topic X be started?") and evaluating 
on-going work ("Based on its track record, will the WG on topic X 
succeed?") ?

A number of formal checks can be identified for both these cases, like 
"is there a critical mass of critical players actively involved?";  "is 
the output of the WG influencing the kind and volume of expectations 
from the intended users?"; "is the WG actually taking advantage of the 
liaisons it has specified in its charter?", etc.

Anyway, I am looking forward to more people adding their views on this 
Success/Failure topic.

/olle

>
> -----Original Message-----
> From: Olle Olsson [mailto:olleo@sics.se]
> Sent: Wednesday, December 10, 2014 8:08 AM
> To: public-success-fail@w3.org
> Subject: What is a failure?
>
> Steve P asked: "definition of success?"
>
> In the same way one can ask: "definition of failure?"
>
> (and of course: "is there a grey zone between success and failure?")
>
> For failure, at least separate between:
>
> (1) failure to reach the formal objectives specified for the work.
> Possibly evaluated at end of chartered period.
>
> (1.a) nothing that was planned to be completed was really finalized. Too tough challenge? To many ideas and no compromises agreed upon?
> (1.b) some deliverables were completed, but without having the whole package of planned deliverables completed, nothing can really be useful.
>
> and
>
> (2) failure to achieve the expected real-world effects. Deliverables completed as planned. Probably evaluated some time after completed work.
>
> (2.a) the targeted usage area seems to have become irrelevant (technology and business moves to other areas)
> (2.b) the targeted usage area is still relevant, but for some reasons the W3C specifications produced have not been adopted in practice, and no other technology solves this problem.
> (2.c) the targeted usage area is still relevant, but at some point in time other technology approaches became more popular, and the W3C technology was marginalized.
>
>
> Identifying failures is more easy when done well into the future (which in internet times could be one year after completed work).
>
> Identifying possible failures early (before work starts, or during work) is difficult. What to look at? Comparing technology approach to use cases? Comparing technology to competing technology (existing or in works)? Is W3C the best place to do the work? Are the critical stakeholders involved? Is there critical mass? Etc. So one can evaluate the technology being worked on, and one can evaluate the work being put into developing technology.
>
> On the page
>      https://www.w3.org/wiki/AB/2014-2015_Priorities/w3c_work_success
> I was intrigued by the list of "Initial list of failures to learn from".
>
> For what reason are the technologies listed there called "failures"? For example, one can these days see quite a lot of enthusiasm about applications of RDF, but nevertheless it is on the failures list.
>
> /olle
>
>
> --
> ------------------------------------------------------------------
> Olle Olsson   olleo@sics.se   Tel: +46 8 633 15 19  Fax: +46 8 751 72 30
>           [Svenska W3C-kontoret: olleo@w3.org] SICS [Swedish Institute of Computer Science] Box 1263 SE - 164 29 Kista Sweden
> ------------------------------------------------------------------
>
>


-- 
------------------------------------------------------------------
Olle Olsson   olleo@sics.se   Tel: +46 8 633 15 19  Fax: +46 8 751 72 30
         [Svenska W3C-kontoret: olleo@w3.org]
SICS [Swedish Institute of Computer Science]
Box 1263
SE - 164 29 Kista
Sweden
------------------------------------------------------------------