On Mon, Apr 25, 2011 at 9:16 AM, Gary Gregory <garydgreg...@gmail.com> wrote:
> On Sat, Apr 23, 2011 at 7:40 PM, Henri Yandell <flame...@gmail.com> wrote:
>
>> On Sat, Apr 23, 2011 at 8:36 AM, Matt Benson <gudnabr...@gmail.com> wrote:
>> > On Sat, Apr 23, 2011 at 6:31 AM, Gary Gregory <garydgreg...@gmail.com>
>> wrote:
>> >> On Apr 23, 2011, at 7:10, "Jörg Schaible" <joerg.schai...@gmx.de>
>> wrote:
>> >>
>> >>> Gary Gregory wrote:
>> >>>
>> >>>> Hi All:
>> >>>>
>> >>>> I find that the new 'valid' method names in Validate make for odd
>> reading.
>> >>>>
>> >>>> I think a verb like 'validate*' or 'check*' would be better.
>> Especially
>> >>>> when the Javadocs all start with 'Validates...'.
>> >>>>
>> >>>> I do see 'check' used in other internal APIs for this kind of
>> behavior.
>> >>>> For example, Java Swing and Eclipse SWT use 'check*' methods to
>> validate
>> >>>> state and throw exceptions.
>> >>>>
>> >>>> For example:
>> >>>>
>> >>>> public void doSomething(String str) {
>> >>>> Validate.validateIndex(str, 1);
>> >>>>
>> >>>> or:
>> >>>>
>> >>>> public void doSomething(String str) {
>> >>>> Validate.checkIndex(str, 1);
>> >>>>
>> >>>> The Validate class name is odd too because it is a verb. I would
>> expect
>> >>>> Validator:
>> >>>>
>> >>>> public void doSomething(String str) {
>> >>>> Validator.validateIndex(str, 1);
>> >>>>
>> >>>> A validator validates (or checks) values.
>> >>>>
>> >>>> I think I like best the 'check*' methods, probably because I've seen
>> them
>> >>>> in SWT and Swing for so long.
>> >>>>
>> >>>> Thoughts?
>> >>>
>> >>> Validator.checkXXX sounds reasonable.
>> >>
>> >> I am moving today, so I might not get to this until later. If someone
>> >> can jump in that would be great.
>> >>
>> >
>> > I think the idea of Validate was to sound "fluently assertive."
>> > "validate that 'some condition'." I don't find this problematic.
>> > Conversely, I do find somewhat problematic the idea that [lang],
>> > arguably the most common of all Commons components, should hijack a
>> > classname that is already central to not only another Commons
>> > component, but also multiple JSRs. Let's be kind to our community and
>> > find another way.
>>
>> Bit harsh - bear in mind that Validate.java was added in December
>> 2002. Validator was brand new having only released a 1.0 in October
>> 2002 and I suspect the multiple JSRs were either non-existent or being
>> typically JSR secretive.
>>
>
> Should this class be dropped then? We do have a whole [validator] project
> after all. Do these methods provide something [validator] does not?
-1 (and I'm not even much of a user of Validate).
They have different use cases; Validator is about validating user
input, Validate is more of an AssertUtils.
Hen
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org
