Sizzle Power in MooTools?

[Aaron]

I've posted a bit about the recent news that Dojo, Prototype, and
others are considering using Sizzle for their CSS engines. What do you
think about MooTools doing the same? I'm curious. Here's my post:

http://www.clientcide.com/industry-news/sizzle-power-in-mootools/

[ibolmo]

Looking into it doesn't mean it's going to happen.

No comment on the unification of frameworks. Too many coding styles
colliding, first of all.

[ibolmo]

Lastly, I am not sure why it's the community's choice to add support
to this. All the pros you mentioned directly affect the development of
the framework/code. I don't see how users will benefit. As you said,
they only see the product.

[nwhite]

If it makes the life of the developers easier, and it makes sense for the framework then ok. I think the speed differences are so small that its not really an issue. Sizzle has some nice optimization but it sounds like Jan is already looking into implementing some of this. My biggest concern would be the loss of extendibility. I have only extended the selector class on one or two occasions but it was a timesaver when I did. I personally would prefer to see Mootools stay pure to its core philosophy and not dilute from its original goals. Losing the option to extend anything for a few microseconds seems like too big of a sacrifice.

just my 2 cents.

[nutron]

It's not the community's choice really. I'm just asking for input.

On Wed, Dec 3, 2008 at 8:34 PM, ibolmo (via Nabble) <ml-user%2B64531-746315738@...> wrote:

Lastly, I am not sure why it's the community's choice to add support
to this. All the pros you mentioned directly affect the development of
the framework/code. I don't see how users will benefit. As you said,
they only see the product.

On Dec 3, 8:31 pm, ibolmo <Olmo.Maldon...@...> wrote:

> Looking into it doesn't mean it's going to happen.
>
> No comment on the unification of frameworks. Too many coding styles
> colliding, first of all.
>
> On Dec 3, 8:22 pm, Aaron <anut...@...> wrote:
>
> > I've posted a bit about the recent news that Dojo, Prototype, and
> > others are considering using Sizzle for their CSS engines. What do you
> > think about MooTools doing the same? I'm curious. Here's my post:
>
> >http://www.clientcide.com/industry-news/sizzle-power-in-mootools/

The MooTools Tutorial: www.mootorial.com CNET Clientside: clientside.cnet.com

---

View this message in context: Re: Sizzle Power in MooTools?
Sent from the MooTools Users mailing list archive at Nabble.com.

[nutron]

This is all great feedback. Thanks. Keep it comin.

On Wed, Dec 3, 2008 at 8:48 PM, nwhite (via Nabble) <ml-user%2B93763-1341009700@...> wrote:

If it makes the life of the developers easier, and it makes sense for the framework then ok. I think the speed differences are so small that its not really an issue. Sizzle has some nice optimization but it sounds like Jan is already looking into implementing some of this. My biggest concern would be the loss of extendibility. I have only extended the selector class on one or two occasions but it was a timesaver when I did. I personally would prefer to see Mootools stay pure to its core philosophy and not dilute from its original goals. Losing the option to extend anything for a few microseconds seems like too big of a sacrifice.

just my 2 cents.

On Wed, Dec 3, 2008 at 11:34 PM, ibolmo <Olmo.Maldonado@...> wrote:

Lastly, I am not sure why it's the community's choice to add support
to this. All the pros you mentioned directly affect the development of
the framework/code. I don't see how users will benefit. As you said,
they only see the product.

On Dec 3, 8:31 pm, ibolmo <Olmo.Maldon...@...> wrote:
> Looking into it doesn't mean it's going to happen.
>
> No comment on the unification of frameworks. Too many coding styles
> colliding, first of all.
>
> On Dec 3, 8:22 pm, Aaron <anut...@...> wrote:
>
> > I've posted a bit about the recent news that Dojo, Prototype, and
> > others are considering using Sizzle for their CSS engines. What do you
> > think about MooTools doing the same? I'm curious. Here's my post:
>
> >http://www.clientcide.com/industry-news/sizzle-power-in-mootools/

[Steve Onnis]

How is it not the communities choice?  If it is not going to suit a lot of people using it then they are going to move away.  Isn't it in the best interest of the project that the community has a voice?

---

From: mootool...@googlegroups.com [mailto:mootool...@googlegroups.com] On Behalf Of nutron
Sent: Thursday, 4 December 2008 3:54 PM
To: mootool...@googlegroups.com
Subject: Re: Sizzle Power in MooTools?

[Garrick]

I decided to post what I thought about using Sizzle:
http://www.garrickcheung.com/mootools/implement-other-selector-engines-eg-sizzle-or-peppy-into-mootools/

[nwhite]

@steve:

Mootools has created it's community from what it is, not from what people want. I honestly don't think having a different selector engine is going to make or break choosing a framework. Aaron pointed out the advantage of sharing with other frameworks we gain when they gain. I see this as an area where there isn't much more to gain. The thing I would hate to see happen is my inability to extend selectors which is a core philosophy of Mootools. I also think that adopting Sizzle would make life harder on the developers. This splinters the core internals of mootools, making re-factoring harder. I think the whole warm and fuzzy feeling of being united with other frameworks especially during the holiday season has some appeal. However, I think it only helps on a PR front, not on a code front. 

One thing that seems to be minimized in all this press is the actual real world usage of selectors. Esoteric selectors are nice to have but are they practical? Usually there is a better way of handling your DOM selection, which leads to better selection performance and readability. The other point is the shear number of selections that are transpiring in these speed tests. No real world app uses selectors so heavily to notice any kind of difference between any framework. Particularly with Sizzle it gains a lot of speed from caching, I cannot recall an instance when I reselect after the initial pass. I almost always handle this in my class. I see selectors as a way of initializing my application, run once to get the pieces I need and keep track of my own state. I believe that Mootools encourages this coding style with the class base structure, further minimizing the speed benefits of caching.

In any event, I trust the developers to choose what is best for Mootools. They have done a terrific job on managing the direction of the framework thus far.

[Tom Occhino]

nwhite, while I agree with most everything you said, and ultimately, Sizzle will not be finding it's way into MooTools, you should note that it is extremely extensible, and things like custom pseudo selectors are no problem at all.  I think I should just make sure everyone is informed before making their arguments ;)

Thanks for your positive feedback, you seem to have a really solid understanding of the forces that have driven the development of MooTools thus far. I feel like you and Valerio would get along really well ;) Hopefully there will be a blog post soon...

Tom

[Aaron]

I've posted a big followup:

http://www.clientcide.com/best-practices/dojos-dylan-schiemann-and-jquerys-john-resig-on-my-sizzle-post/

> > From: mootool...@googlegroups.com [mailto:mootool...@googlegroups.com
> > ] On Behalf Of nutron
> > Sent: Thursday, 4 December 2008 3:54 PM
> > To: mootool...@googlegroups.com
> > Subject: Re: Sizzle Power in MooTools?
>
> > It's not the community's choice really. I'm just asking for input.
>
> > On Wed, Dec 3, 2008 at 8:34 PM, ibolmo (via Nabble) <ml-user
> > %2B64531-746315738@...> wrote:
>
> > Lastly, I am not sure why it's the community's choice to add support
> > to this. All the pros you mentioned directly affect the development of
> > the framework/code. I don't see how users will benefit. As you said,
> > they only see the product.
>
> > On Dec 3, 8:31 pm, ibolmo <Olmo.Maldon...@...> wrote:
>
> > > Looking into it doesn't mean it's going to happen.
>
> > > No comment on the unification of frameworks. Too many coding styles
> > > colliding, first of all.
>
> > > On Dec 3, 8:22 pm, Aaron <anut...@...> wrote:
>
> > > > I've posted a bit about the recent news that Dojo, Prototype, and
> > > > others are considering using Sizzle for their CSS engines. What  
> > do you
> > > > think about MooTools doing the same? I'm curious. Here's my post:
>
> > > >http://www.clientcide.com/industry-news/sizzle-power-in-mootools/
>
> > The MooTools Tutorial:www.mootorial.comCNET Clientside:  
> > clientside.cnet.com
>

[keif]

Honestly, I'm torn, but ultimately it's heading away from the old
school bragging contest these framework discussions used to be.

We're splitting hairs with CSS selectors - we're done to milliseconds.
Break out a stop watch and time it yourself - can you really tell
which code is faster, watching it? Doubtful.

I do echo the sentiment that it would *just be so damn cool* if we
adopted it - for the sole reason of "we can all get along" - but if
it's not heading down a path of where I can drop in any framework and
run with my code, than the whole "buddy buddy" feeling wears off. And
then what about when the time comes when Developer A feels like he's
made a break through? Maybe Aaron rewrites part of it and John
disagrees (over simplifying my case) so now mootools is slightly
different than jquery, and we're back to where we were - combining
great ideas into our respective frameworks.

The idea *is* cool - but cool ideas don't necessarily need to happen.

I'd like MooTools Developers and users to check out sizzle and decide
for themselves - and those that work in the core of MooTools to
compare the code and decide if it's worth it to switch, or to learn
from it what they can to make mootools better (and separate). There's
no harm in saying "thanks for the opportunity, we'l take it to heart
and incorporate some of the ideals (maybe)."

Just my quick thought.

-keif

On Dec 4, 1:58 pm, Aaron <anut...@gmail.com> wrote:
> I've posted a big followup:
>
> http://www.clientcide.com/best-practices/dojos-dylan-schiemann-and-jq...

[Jan Kassens]

We had a good discussion about Sizzle yesterday and decided against
it. Valerio summed up some of the reasons we don't include Sizzle now
and likely won't do that in the future: http://mootools.net/blog/2008/12/04/sizzle/

--
Jan - MooTools comitter
twitter/blog: http://kassens.net