Re: ISSUE-76: Need feedback on splitting Microdata into separate

Ian Hickson, Thu, 10 Dec 2009 16:12:34 +0000 (UTC):
> On Thu, 10 Dec 2009, Leif Halvard Silli wrote:
>> Ian Hickson, Thu, 10 Dec 2009 14:41:03 +0000 (UTC):
>>> On Thu, 10 Dec 2009, Leif Halvard Silli wrote:
>>>> 
>>>> But Microdata is *already* being designed by only a half or a third 
>>>> of the group, despite that it is placed inside the same spec. This 
>>>> split will be/is already reflected in the design.
>>> 
>>> Could you point to where microdata's design mirrors the communication 
>>> structure of the working group?
>> 
>> Why should I need to explain the effect of Conway's law anymore than Tab 
>> does? Does Conway even explain how? Otherwise, I stick to the quote from 
>> Wikipedia as the definition of what Conway's laws is: a "valid 
>> sociological observation". The HTML 5 work has a structure. A strange 
>> structure: Two groups instead of one. From early on, I said that the 
>> logical thing would be if WHATwg went into hiatus (at least w.r.t. HTML 
>> 5), and instead recommended that all activity took place in this group. 
>> That is the _true_ solution to the problem that Conway's law describes. 
>> As it is, Microdata _especially_ is an example of something that 
>> happened in WHATwg instead of in this WG. But if you think that _that_ 
>> neither was or is a problem, then all the less should it be a problem 
>> for Microdata to be placed in another spec, within the same two-group 
>> structure.
> 
> That didn't answer the question.

I beg to disagree. But see below.

> When talking about microdata, you said that the split that Conway's Law 
> predicts "will be/is already reflected in the design". Could you point to 
> where microdata's design mirrors the communication structure of the 
> working group?
> 
> Tab was saying that there is no such split, but that splitting it out 
> would lead to such a split. I understand how that could happen -- if 
> different parts of a technology are designed by independent and separate 
> groups not closely working together, then Conway's Law applies. However, 
> you said that it either had already happened, or would happen, presumably 
> even without such a split. I'm just asking for you to point to where this 
> split is, to determine what you consider to be such a split.
> 
> Presumably it should be easy enough to point to such a split, since if 
> there wasn't such a split, you wouldn't have suggested there was.

There is no agreement within this group that we are working closely 
together on Microdata, which is what Conway's mandates in order avoid 
the effect of the law. Conway's law says that under such circumstances, 
things will not end up compatible. I don't think that Conway's law says 
that you will always - immediately - be able to pinpoint the 
incompatibilities. But I read Conway's law to say that they will 
nevertheless be there.

In a way, Microdata is in itself an example of Conway's law: A parallel 
system to something that already exists. From Wikipedia:

> government hires company X to build system S. Say company X has three 
> engineering groups, E1, E2, and E3, that participate in the project. 
> Conway's law suggests that it is likely that the resultant system 
> will consist of 3 major subsystems (S1, S2, S3), each built by one of 
> the engineering groups.

Story: W3 hired the HTMLwg and the WHATwg to make HTML 5 ...

Much of the communication within our two-group system has been about 
avoiding that Microdata steps RDFa needlessly on its toes (avoiding 
using the same attribute etc.) But even so, I have observed claims from 
the Microformats community about repeating their errors. 

Since you dismissed what I said about that WHATwg should go into 
hiatus, I feel tempted to quote Wikipedia again:

> organizations that are not willing to re-organize to generate an 
> optimal design, can end up producing a sub-standard design that 
> merely reflects the pre-existing organization.
> But the essence of Conway's Law also applies to flexible 
> organizations that are willing to re-organize to produce an optimal 
> design.

I think that you will have more support from the critical groups if 
moved to a separate spec. That, in turn, should be a move that could 
widen the organisation/structure/process around Microdata. That in turn 
should ensure a better - more compatible - outcome. The current work on 
Microdata has not had wide support within this group. And perhaps even 
less outside this group. I can't see that this could worsen by being 
moved into another spec. And hence, Tab's premise is wrong.
-- 
leif halvard silli

Received on Thursday, 10 December 2009 16:54:32 UTC