Re: Design Principles, Section 1.6.1 relationship to HTML 4.01

Hi Leif,

I hesitate to drag out this discussion. I note that at various times  
I've gotten the following requests:

- Add disclaimers or clarifications to principles where it would be  
wrong to apply them as absolute 100% of the time.
- Remove disclaimers from principles because it is wrong to imply  
there may be exceptions.
- For any principle with exceptions or disclaimers, either state is as  
absolute or remove the principle if it cannot be.
- State principles as absolute even if they would seemingly conflict  
or lead to wrong answers if taken as such; assume balancing of  
competing concerns is external.
- Place an overall giant disclaimer on the HTML Design Principles  
document as a whole.

So, clearly, you can't please everyone. But given the feedback pushing  
in opposite directions (sometimes even from the same people at  
different times), I am inclined to think the current level of firmness  
is about right. As long as the principles are not absolute to the  
point of clear absurdity, or so couched in disclaimers that they are  
weasel-worded fluff, then I see no cause for concern.

Further, I don't think a lengthy discussion of absoluteness vs.  
disclaimers would be very productive. In my opinion, the ideas  
encapsulated in the principles are more useful and interesting than  
the details of their wording. Admittedly, these ideas are not as  
useful as a concrete specification. But they have given the group some  
shared vocabulary to discuss various ideas, even if we do not always  
ultimately agree.

Regards,
Maciej


On May 31, 2009, at 5:46 PM, Leif Halvard Silli wrote:

>> I guess we disagree on the basics of proper principles for language  
>> design.
>> If I understand your position correctly, you're saying any rule  
>> that can't be stated and obeyed in 100% absolute terms is not worth  
>> writing down.
>
> Ahem ...
>
> Principles are principles. Stating clear and unambigous principles  
> is one thing. Finding, that in reality, you have to break a  
> principle or two, is another thing.
>
>
>> I disagree.
>
> Who are you arguing with - a strawman?
>
>> Software engineering design is a world of tradeoffs. Often design  
>> problems involve balancing multiple competing factors. But that  
>> makes it important to be clear about your design goals, and then  
>> apply appropriate judgment. It does not mean throw out your design  
>> goals. And it does not mean make them absolute rules so the process  
>> can work without human judgment.
>
> One can have multiple competing principles. /Then/ we can do the  
> tradeoff. You do not need to express all this judgement within the  
> principles. Doing /that/ is poor judgement.
>
> Otherwise, I note that you mention "goals".
> -- 
> leif halvard silli
>

Received on Monday, 1 June 2009 13:18:43 UTC