Peter, I agree with you on the BigDecimals. Get it working with the right functionality, and refactor the implementation later.
What do we need to do with the numerics patch to have it be "good enough" to submit officially for CVS access? It'd be nice if the addition was comprehensive enough that it could be immediately used without a lot of caveats. I suggest -- basic arithmetic & logic (done) -- TemplateNumber (done) -- Comparable (probably not hard) -- decimal literals Anything else? When this is ready I suggest we officially request Geir to patch the CVS so a wider audience can play with this. I know we're all waiting to see what happens then. :-) Best, WILL ----- Original Message ----- From: "Peter Romianowski" <[EMAIL PROTECTED]> To: "Velocity Developers List" <[EMAIL PROTECTED]> Sent: Thursday, June 26, 2003 1:26 PM Subject: Re: Velocity Numerics [PATCH] > Nathan Bubna wrote: > > [snip] > > > still, i'd like to see us focus first on the features that we know people > > want. i'm afraid we might bite off more than we can chew. Geir says he'll be > > That's exactly why I wanted to cut the Numbers and String comparison thing. > > > a patch-horse at least until we get a new committer in there. let's not miss > > this opportunity to get some of these "big" things (floats, maps, whitespace) > > numbers, maps and *filters* ;) > > > in by getting mired down in "little" ones. we have some energy and momentum > > now, let's keep it focused. > > You're right. BTW I "miss" Mat�as Giovannini a bit who had some good points > about the implementation details of the whole numbers-thing. At this time > the patch is completely based on BigDecimal which might be a real performance > drawback in some situations. I'd like to stick to that version until it > gets committed (to the DEV-branch of course) along with the extensions Will made > (and others to come) and then improve it. The reason is simple: New features > around the number thing can be added quickly and we are more focused on the > solution. When the numbers thing has been proven to work correct for all then > I'd like to work on a faster implementation which should be quite easy to do. > > (I have to admit that I don't want to put too much work into it since I already > did and it lead to nowhere then. I'd just like to see the 'something' happening ;) > > Peter > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
