Not sure if this matters for your application but caching the view seems like a good idea if you use this method. I would think there would be some performance hit by doing the file system read. Just something that I was noodling on when re-reading your article.

Thanks again for your thoughts.

Also, even for me I don’t use a separate JS/CSS file for every controller/action, but because the view helper only loads the files *when* they exist that works perfectly fine.

http://www.devpatch.com/2010/02/view-plugin/

Thoughts and opinions are welcome.

My thoughts for a packer/minifier would probably be to take hints from this project and this one (if not just using them outright). Instead of including the file, read the contents of the file and send it to the packer/minifier class. Send the returned value to a Zend_Cache object. The view helper could handle the caching and either output the script inline (which would be the best bandwidth optimizing, however, I suspect you might run into some CSS rendering issues) or include it all in one simple file.

If I have time, I’ll try to whip this up — but in my experience thus far, this is more of a nice-to-have than a requirement. When I first started using my CSS/JS view helpers I was very concerned about loading an additional file per controller/action, but in practice I’ve discovered the overhead is pretty minimal for the maintainability you end up getting out of it.

Even better, I am relieved that you, too, use page-specific (in the case of an MVC-framework like Zend Framework, controller-action-specific) styles and scripts. Looking at so many sites and samples, it was beginning to appear that best practice was a single global css file and a single javascript file (to reduce the number of HTTP requests, etc). But it always bugged me that there would be some styles/scripts that would be page-specific, so why should they be parsed/executed on *every* page?

Thanks for a bit validation that at least I am not alone in this. On the other hand, any ideas on the best way to address this – like your allusion to packing/minifying in a way that would still create significant caching benefits – would be most welcome.