Subject: Re: [boost] [guidelines] why template errors suck
From: Robert Ramey (ramey_at_[hidden])
Date: 2010-10-04 15:08:16
David Abrahams wrote:
> At Mon, 4 Oct 2010 10:04:04 -0800,
>> Maybe my reservations about the idea of "formal semantics" are more
> They're clear. You are saying, AFAICT, that you don't want to try to
> pin down any semantics because that might rule out, purely on the
> basis of specification, some useful application of your existing code.
> The only problem with that is that without some such specification,
> nobody even knows if their existing applications of your code are
> going to work tomorrow. We're in a "use the source, Luke" situation
> at the moment.
That's part of it. But its more than that. I have no problem with saying
that the library does this or that. In fact the documentation for the
does define what the word means in the context of that documentation.
It defines it in english appealing to shared intuition and world view
between the writer
and reader of the document. I doesn't do so in any formal sense. I
don't think that is possible and that attempts to do so just end up
requiring a sharing of other ideas. So it's better just to cut the
recurrsion and just accept that we share and idea of what serialization
and de-serialization are but that idea can have differing interpretations.
To summarize - I'm just very uncomfortable with the term
"formal symatics" and don't think it can mean anything.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk