Peter gummer at 21:43, 13 February 2008

2008-02-13T21:43:26Z

← Older revision		Revision as of 21:43, 13 February 2008
Line 2:		Line 2:
	Although this point is tangential, and doesn't really affect the problem and solution that Alexander describes, I wonder whether it was such a good idea to make ''note'' a keyword. Few Eiffel keywords are likely to be wanted as identifiers by application developers. The obsolete ''indexing'' keyword, although pointlessly verbose, did have the virtue that no one was likely to want to declare ''indexing'' as an identifier. Not so for ''note''; ten years ago I was working on a system (in Delphi, not Eiffel) where "note" was a central concept.		Although this point is tangential, and doesn't really affect the problem and solution that Alexander describes, I wonder whether it was such a good idea to make ''note'' a keyword. Few Eiffel keywords are likely to be wanted as identifiers by application developers. The obsolete ''indexing'' keyword, although pointlessly verbose, did have the virtue that no one was likely to want to declare ''indexing'' as an identifier. Not so for ''note''; ten years ago I was working on a system (in Delphi, not Eiffel) where "note" was a central concept.

	Personally, I think I would have preferred to replace "indexing" with some kind of punctuation. C# ''attributes'', which perform a pretty similar role to Eiffel ''notes'', have a syntax that I have always found perfectly legible. Eiffel tends to prefer keywords over punctuation, in order to be as legible as possible for non-geeks; but not completely: for example, in Eiffel we write '':='' rather than ''compute'', ''--'' rather than "rem", and the new object test syntax is about as far from intuitive as anything I've seen in any language.		Personally, I think I would have preferred to replace ''indexing'' with some kind of punctuation. C# ''attributes'', which perform a pretty similar role to Eiffel ''notes'', have a syntax that I have always found perfectly legible. Eiffel tends to prefer keywords over punctuation, in order to be as legible as possible for non-geeks; but not completely: for example, in Eiffel we write '':='' rather than ''compute'', ''--'' rather than "rem", and the new object test syntax is about as far from intuitive as anything I've seen in any language.

	The new ''note'' syntax isn't in EiffelStudio 6.1 yet. Is it too late to reconsider?		The new ''note'' syntax isn't in EiffelStudio 6.1 yet. Is it too late to reconsider?

	----		----

Peter gummer at 21:42, 13 February 2008

2008-02-13T21:42:25Z

New page

--[[User:Peter gummer|Peter gummer]] 13:42, 13 February 2008 (PST)
Although this point is tangential, and doesn't really affect the problem and solution that Alexander describes, I wonder whether it was such a good idea to make ''note'' a keyword. Few Eiffel keywords are likely to be wanted as identifiers by application developers. The obsolete ''indexing'' keyword, although pointlessly verbose, did have the virtue that no one was likely to want to declare ''indexing'' as an identifier. Not so for ''note''; ten years ago I was working on a system (in Delphi, not Eiffel) where "note" was a central concept.

Personally, I think I would have preferred to replace "indexing" with some kind of punctuation. C# ''attributes'', which perform a pretty similar role to Eiffel ''notes'', have a syntax that I have always found perfectly legible. Eiffel tends to prefer keywords over punctuation, in order to be as legible as possible for non-geeks; but not completely: for example, in Eiffel we write '':='' rather than ''compute'', ''--'' rather than "rem", and the new object test syntax is about as far from intuitive as anything I've seen in any language.

The new ''note'' syntax isn't in EiffelStudio 6.1 yet. Is it too late to reconsider?

----

Talk:Syntax level - Revision history

Peter gummer at 21:43, 13 February 2008

Peter gummer at 21:42, 13 February 2008