Naming Guideline Discussion

TimeZones

Community Blogs — Sat, 07 Oct 2006 05:08:52 GMT

Right now, there is no easy way to convert a time from one arbitrary timezone to another arbitrary timezone

re: Naming Guideline Discussion

Dave Griffiths — Sun, 08 Oct 2006 13:07:29 GMT

Whats wrong with TimeZoneInfo? .. short, sweet, keeps to the same naming convention as RegionInfo, CultureInfo .. "insert any other typename here"Info Using a 2 seems daft and goes completely against any class naming conventions .. thought the primary goal of naming something is that the name is self-describing of what it does/contains. TimeZone2 just says that it's version 2 of TimeZone. And that hasn't been done in many other places in the BCL (with maybe the exception of a few interface names)

Should Humans Be Called Ape2?

Omer van Kloeten's .NET Zen — Tue, 10 Oct 2006 23:21:26 GMT

Ever since Kathy Kam announced on her weblog that a new type named TimeZone2 will be introduced into

re: Naming Guideline Discussion

diegocanepa — Wed, 11 Oct 2006 01:12:23 GMT

What about TimeRegion?

Regards

re: Naming Guideline Discussion

Matt Ellis — Wed, 11 Oct 2006 15:01:44 GMT

Hi Kathy. Interesting topic. I've got a suggestion to sidestep the whole -2 suffix completely. This new class doesn't really add any new functionality to TimeZone. The TimeZone class describes an actual time zone, and (as I understand it) the new functionality is to return an instance of TimeZone based on an id (factory pattern), get a list of known time zones and to convert from one timezone to another.

Creating a TimeZoneFactory class could solve both the first and second problems and a TimeZoneConverter class would solve the other one . Any other methods that add functionality to TimeZone itself can be added on - adding methods/properties isn't a breaking change. A method on TimeZone could be used to convert a date from one timezone to another.

This solves your versioning problem, and also keeps the responsibility of TimeZone(2) to a single responsibility.

What do you think? And conversely, what was the thinking behind adding these extra responsibilities to TimeZone(2)?

Cheers!

Matt

re: Naming Guideline Discussion

KathyKam — Thu, 12 Oct 2006 00:34:07 GMT

Hi Matt,

Thanks for the suggestion! However, this is incorrect...

"Any other methods that add functionality to TimeZone itself can be added on - adding methods/properties isn't a breaking change."

Actually, we changed the behaviour of methods like IsDaylightSavingTime. Read Part 2 of my "Desiginng TimeZone2" for more detail. If you have more questions or comments after that. Please let me know.

Regarding adding these extra responsibilities to TimeZone2. Once we realize we need a new type regardless, we need to make a decision on how many new types we should add. We decided to add one. Usability studies shows that it is easier for our users to deal with only one type for all their need. That said, I want to emphsis that API design is all about tradeoffs. No one design is perfect, and there are of course arguments to say 2 (or even 3) types are better. I'd love to know why you think 2 or even 3 types are better than 1!

Thanks,

Kathy

re: Naming Guideline Discussion

David M. Kean — Fri, 13 Oct 2006 07:48:45 GMT

When you see TimeZone and TimeZoneInfo next to each other in intellisense, how do you know which one to use?

If you say TimeZone and TimeZone2 next to each, I think it is a lot more clearer that TimeZone2 is the later version.

re: Naming Guideline Discussion

KathyKam — Wed, 18 Oct 2006 02:58:27 GMT

Glad you think TimeZone2 is better! :) I wonder does anyone else thinks it is better.

re: Naming Guideline Discussion

diegov — Fri, 24 Nov 2006 17:03:46 GMT

Kathy, I read the guidelines and I think the TimeZone2 name is the least harmful ;)

Anyway, what I don't find explained in detail is why you had to create a new type instead of "stuffing" the new features in the original class (and possibly marking part of its remains as obsolete). I am sure it was because you found some road block there and not because of a matter of taste. What was it?

re: Naming Guideline Discussion

diegov — Fri, 24 Nov 2006 17:06:41 GMT

Ah, now I read your comment. So you changed behaviors... I will read the rest in your "Designing TimeZone2^2" ;). Thanks!

re: Naming Guideline Discussion

Ben Siêu Nhân — Wed, 03 Jan 2007 21:23:37 GMT

@David M. Kean & Kathy: if you insist that TimeZone2 beats TimeZoneInfo in indicating it is a "new" one, how about putting Obsolete attribute to TimeZone to imply that it is an "old" version of TimeZoneInfo?

Just imagine some day they want to "upgrade" such things as Int32... there you go, we got Int322 !!! Now, tell the innocent why it is "Int32 version 2" but not "Int version 322" ???