REBOL3 tracker
  0.9.12 beta
Ticket #0002155 User: anonymous

Project:

Previous Next
rss
TypeWish Statusproblem Date15-Jun-2014 20:22
Versionr3 master CategorySyntax Submitted byfork
PlatformAll Severityminor Prioritynormal

Summary Make comma and period equivalent for symbol lookup in words
Description There is a purposeful Rebol philosophy of making anywhere a comma can appear be equivalent to a period.

One reason is that for reading purposes, being able to see the difference between a comma and a period is hard. Making them have linguistic significance has bad ergonomics. A useful side-effect is compatibility with countries where commas are used instead of periods as a decimal point (as we saw in Montreal).

The comma is not currently allowed in symbols. But it seems that much like how case is preserved (but not distinguished) a comma could be preserved but not affect the hashing of the symbol vs a period. Under this thinking:

>> 'ABC,DEF,GHI = 'abc.def.ghi
== true

>> mold 'ABC,DEF,GHI
== "ABC,DEF,GHI"

>> mold 'abc.def.ghi
== "abc.def.ghi"

A dialect which wished to distinguish between comma and period inside of a symbol would be on its own with that...just as a dialect today that wants to discern case must do.
Example code

			

Assigned ton/a Fixed in- Last Update22-Jun-2014 22:00


Comments
(0004459)
BrianH
22-Jun-2014 02:25

Too difficult to visually distinguish a comma and a period, for something that would have to be considered distinct with the stricter equivalences. This isn't like case, it's easier to distinguish case differences visually (with the exception of l vs. I vs. 1, or O vs. 0). This would feed into the aversion the Rebol community has about the comma, and wouldn't add any benefit to offset this (who would logically consider a comma and a period equivalent?).
(0004469)
fork
22-Jun-2014 22:00

The benefit is the ability to use comma if you prefer it to period for some purpose (I've named a purpose where I prefer it, though am also willing to tolerate period if one wished).

Lack of distinguishability is not a problem in almost all circumstances. The only cases that would be concerned here would be dialects that are doing their own symbolic lookup outside of the Rebol system, and they may be handcuffed by legacy concerns to worry over the distinction... just as they may be concerned about case.

Date User Field Action Change
22-Jun-2014 22:00 Fork Comment : 0004469 Modified -
22-Jun-2014 22:00 Fork Comment : 0004469 Added -
22-Jun-2014 02:25 BrianH Description Modified -
22-Jun-2014 02:25 BrianH Status Modified submitted => problem
22-Jun-2014 02:25 BrianH Comment : 0004459 Added -
15-Jun-2014 20:22 fork Ticket Added -