ibatis-user-java mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Clinton Begin <clinton.be...@gmail.com>
Subject Re: [HELP] I can't not type < sign in my sqlmap xml file
Date Tue, 24 May 2005 19:37:52 GMT
I agree with Brandon. There are definite improvements we need to make to 
iBATIS dynamic SQL, but that does NOT include building a full templating 
language. The power of iBATIS dynamic SQL is the fact that it is VERY 
specific to SQL. Using iBATIS dynamic SQL, you can reduce the amount of code 
to less than Java, Velocity or Freemarker. All of the latter solutions will 
result in very similar logic, with different syntax. iBATIS replaces common 
logic with smart tags that know what SQL is.

Cheers,
Clinton


On 5/24/05, Brandon Goodin <brandon.goodin@gmail.com> wrote:
> 
> "I guess it's always been a bit of a merry go round."
> 
> This may be true in cases where folks allow thing to get out of hand.
> IBatis has maintained a very focused approach. There has not been a
> real need to introduce a full blown templating language. Cuz if your
> requirements get that complex... do it in java. That is our
> philosophy. Freemarker will suffer from the same problem in ibatis. it
> introduces too much complexity. I think templating languages like
> Velocity and FreeMarker are great stuff. But, IBatis does not require
> or benefit from their robustness. But, in a future version (when we
> aren't XML dependent) folks may have an opportunity to prove me
> wrong... and that is MORE than okay with me :).
> 
> Brandon
> 
> On 5/24/05, Huy <huymail@swiftdsl.com.au> wrote:
> > Brandon Goodin wrote:
> > > I don't plan on promoting any templating languages. It offloads the
> > > burden to another layer. The dyanmic conditionals would get out of
> > > hand. At that point it would be far easier to handle things in java.
> > > The heirarchal nature of xml is of great importance to dynamic sql.
> > > So, i still see it as being the easiest means to manage the SQL. If
> > > you are using a template language you already violate the verasity of
> > > your sql outside of ibatis. So, &lt; is a minor issue IMO.
> >
> > I know it's not a huge deal, but seeing my queries use alot of &lt; and
> > &gt; and not to mention &lt;= i find it adds an extra 0.5 sec each time
> > i read my sql for my brain to do the conversion. This of course
> > doesn't seem like much time but it gets to you after a while.
> >
> > I guess it's always been a bit of a merry go round. First the sql is
> > strung together in java with all the power of a programming language,
> > then it's moved out to a separate template because the code got to
> > messy. Then we start adding simple programming constructs into the
> > template because we needed the flexibility. We keep adding constructs
> > until oops the template starts looking like a programming language and
> > we start moving some of our logic back into our java code.
> >
> > This kinda sounds like HTML templating and it's always been a measure of
> > taste vs complexity when people decide on their templating solution of
> > choice. I'm not sure what my point is, but I see the ibatis
> > semi-templating of SQL will only increase in complexity as the user base
> > grows.
> >
> > > In version 3.0 one of our goals is to remove the dependency on the
> > > configuration being solely XML and allow for java based configuraiton.
> > > This would mean that the xml would be a way to implement the
> > > configuration. XML will continue to be the primary supported means of
> > > configuration because there are so many reasons to continue in this
> > > direction versus other templating languages.
> > >
> > > But, in the future if you so desire, it may be possible for you to
> > > contribute something cool with velocity. Personally, I don't think
> > > velocity would add value to the process.
> > >
> >
> > Maybe you're right. I'll look into Freemarker :-)
> >
> > Huy
> >
>

Mime
View raw message