Just Launched: You can now import projects and releases from Google Code onto SourceForge
We are excited to release new functionality to enable a 1-click import from Google Code onto the Allura platform on SourceForge. You can import tickets, wikis, source, releases, and more with a few simple steps. Read More
Tom Sawyer [mailto:transami@...] wrote:
> / \
> / \
> \/ \/
> logical funtional
> how's that for simplifying things?
Whenever I get tangled by this sort of thorny issue I tend to want to revert
to the physics of the thing. The syntax is very much "physics", and we all
seem to agree on that. The open question that started all this debate was
what operations al I allowed to do on this text file - specifically, can I
or can't I change the orders of the keys of a normal map without breaking
any YAML application that reads it. I say "yes you can safely reorder" and
expect people to use '- ' to signal that this operation may not be done. I
think Mike captured this best.
The models, when all is said and done, provide constrains on what such
"physical" operations preserve semantics and what operations do not. Of
course there are inherently two (and just two) levels here. S-blind and
So about equality; for an S-blind tool changing '010' to '8' breaks
equality; equality can only be "fully" supported by an S-aware tool (at
least partially aware anyway). However an S-blind tool can provide an
approximation of equality, which if good enough in most cases. I agree with
your analysis that, e.g., style-equality is hardly useful.
I _think_ your post ends up correlating to this... If so, fine :-)
From: Clark C. Evans <cce@cl...> - 2002-09-13 15:43:40
On Fri, Sep 13, 2002 at 12:20:24PM +0300, Oren Ben-Kiki wrote:
| So about equality; for an S-blind tool changing '010' to '8' breaks
| equality; equality can only be "fully" supported by an S-aware tool (at
| least partially aware anyway). However an S-blind tool can provide an
| approximation of equality, which if good enough in most cases. I agree with
| your analysis that, e.g., style-equality is hardly useful.
Ahh. This is an interesting perspective on equality. At the non-typed
level; we can use equality of ( !type-identifier, content ); and then
at the typed level we have a type-defined equality which can be _more_
restrictive but not less restrictive. That is, at a lower level we
can detect duplicates based on string equality; but perhaps not all
of the duplicates which would require type knowledge.