Re: API Change Proposal: Change UDTS_WINDOWS_DATE_TIME to UDTS_DOTNET_DATE_TIME

To add to George's answer for #1, the name UDTS_WINDOWS_DATE_TIME is 
just plain wrong. Since this API is brand new, there doesn't seem to be 
much reason to keep the old name around. I think that it would just 
cause confusion.

To answer #2, it seems to me that the .NET platform is key to everything 
that Micorsoft is doing for the next few years, and is not likely to go 
away anytime soon. If they *do* change the name, we can always add an 
alias with the new name.

Regards,
Eric

George Rhoten wrote:
> To answer #1, the proposal states that it's to change the enum.  Since 
> this is still draft API, there is no reason to keep the old constant 
> around.  This is not stable API that is being proposed for change.
> 
> George Rhoten
> IBM Globalization Center of Competency/ICU  San José, CA, USA
> 
> 
> 
> Tex Texin <[email protected]> 
> Sent by: [email protected]
> 01/23/2005 03:07 AM
> 
> To
> Eric Mader <[email protected]>
> cc
> [email protected]
> Subject
> Re: API Change Proposal: Change UDTS_WINDOWS_DATE_TIME to 
> UDTS_NOTNET_DATE_TIME
> 
> 
> 
> 
> 
> 
> 1) Is the proposal to change the name, or to add the new name, while
> keeping the old name for compatibility, and consider the old one
> deprecated?
> 
> 2) Is it a good idea to use names so closely tied to a product?
> These products change names and get reinvented every couple of years,
> although the data types are much more stable. (com, dcom, ole, ocx...)
> 
> Would it be better to use a name that reflected the 64 bit nature and
> the offset (1/1/1/) of the datatype and then separately document this
> data type is used by dotnet?
> 
> Then when the product name changes, the API doesn't.
> The name can then also be shared if the data type is used by other
> non-windows software, instead of adding new constant names for the same
> thing.
> 
> tex
> 
> Eric Mader wrote:
> 
>>Please send comments by 01/28/2005.
>>
>>The universal time scale API in utmscale.h and
>>com.ibm.icu.util.UniversalTimeScale.java talks about a time scale called
>>"Windows DateTime." There isn't any such data type of time scale in the
>>Windows 32-bit API. What was meant was .NET's System.DateTime structure.
>>
>>This proposal is to change the names of the constants that reference
>>this time scale to correctly reflect it's origin:
>>
>>ICU4C: UDTS_WINDOWS_DATE_TIME => UDTS_DOTNET_DATE_TIME
>>ICU4J: WINDOWS_DATE_TIME => DOTNET_DATE_TIME.
>>
>>The actual data for this time scale (a 64-bit integer of ticks since
>>Midnight January 01, 0001) is correct; we're only changing the name to
>>more correctly reflect the source of the time scale.
>>
>>Regards,
>>Eric Mader
>>IBM GCoC - San José
>>5600 Cottle Road  M/S 50-2/B11
>>San Jose, CA 95193
>>
>>_______________________________________________
>>icu mailing list
>>[email protected]
>>http://oss.software.ibm.com/developerworks/oss/mailman/listinfo/icu
> 
>