Re: comments on draft-ietf-sming-reqs-03.txt

[Jon Saperia <saperia@jdscons.com>]

> Hi!
> 
> Jon> [...] If I thought that adding semicolons at the end of
> Jon> each line or changing the capitolization rules would advance the
> Jon> state of management, I would be be all for them. I just have not
> Jon> been convinced that they will.
> 
> This doesn't belong directly to the requirements discussion, but I'd
> like to explain why we introduced the semicolon (back early in the
> NMRG SMIng work).
> 
> We had the goals in mind to
> 
> - design the language so that it's human readable (in the `human ==
>   programmer' sense, that might be a wrong assumption ;-)),
> 
> - design the language so that it's easy to implement parsers, since
>   today we suffer a lot from a variety of bad parsers,
> 
> - design the language so that it's extensible in a way that language
>   extensions (either standardized some day, or vendor specific, or
>   user specific) are parsable or can be skipped if unknown.
> 
> These three goals are listed in the current requirements draft.
> 
As I mentioned earlier. I have heard these points before and simply do not 
agree with them. Optimization for parsers is probably the wrong priority. 
Parsers, and their availability has not been, and does not appear now to be a 
problem in the community.

> We found that the general concept of 
> 
> - statements,
> - statements being introduced by a keyword,
> - statements being terminated by a semicolon,
> - 0, 1, or more arguments to the statement inbetween the keyword and ;,
> - the possibility to let an argument be another sequence of statements
> 
> is very familiar to many people, e.g., from programming languages like
> C and Java. It's a question of how much we are willing to free our
> minds from ASN.1 and open them to what people are familiar with *in
> general*.
>
> The same is true for writing keywords with lowercase letters in most
> computer languages.
> 
>  -frank
> 
Frank, I have come to really like Java. In fact I would object less if we were 
to have tossed everything in favor on Java (I am not seriously suggesting 
that). As you know Java uses a semicolon quite a lot (but not at the end of 
every line). My point is that such changes (like the semicolon) do not gain us 
much. My mind is free - I love and use lots of semicolons in Java. :-) I just 
remain unconvinced that the changes you propose will have the benefits you 
believe. Until there is hard evidence that this is a problem worth solving I 
would leave things alone. That said, that does not mean that we could not make 
improvements to the SMI. We have discussed some items on which I think 
agreement could be reached such as aggregate objects. Those are interesting to 
me and do not require the types of changes that we discussed in this message.

I do not plan on spending a lot of time on these issues. I just wanted to be a 
good citizen an express my views one more time.

/jon
Thanks,
/jon
--

Jon Saperia		     saperia@jdscons.com
			     Phone: 617-744-1079
			     Fax:   617-249-0874
			     http://www.jdscons.com/