Re: html 5 performance suggestion.

> <archive src=”archive_of_css_and_js.gz” alt=”These files are static.”>
> <archive src=”archive_of_images.gz”        alt=”These files are dynamically changing often”>

> When these files were needed they would already be in the browser cache.
> <script src=”abc.js”
> <img src=”abc.gif”
> <video src="abc.mp4"
>  

I like the idea.
But, it seems like you'd want an id on the archives and then be able to specify an archive on the varous tags:
<archive id='scripts' src="archive_of_css_and_js.gz" alt='These files are static" id="scripts" />
<script src='abc.js' archive='scripts' />
...

It would also be nice to be able to package a multi-view 'page' as a primary document with an archive of html template files for use with something like EJS (http://embeddedjs.com). Meaning JS access to archive content would be helpful:
<archive id='page_sub_views' src="subpages.gz" alt='page sub views' />
<script>
var subview= document.archives.page_sub_views.text;
</script>

Then JQuery or one's framwork of choice could be used to append or replace the contents of a div or other element with the contents of subview.

As Peter suggested it means one could front-load content for caching and reduce round trips to the server.

Art C.

Arthur Clifford

That would allow you to have more than one archive with an abc.gif file in it. 
On Nov 28, 2013, at 9:14 AM, "Peter Lenahan" <peter_lenahan@verizon.net> wrote:

> HTML 5 introduced the Manifest feature for offline applications.
> However the content delivery for the items in the Manifest are still single URL references.
>  
> Have you ever considered that web pages are really like Multi-part e-mails.
> They consist of many files to render. Yet the HTML language has no facility to request multiple items in a single http request.
>  
> I would like to suggest that a new tag or attribute be added to the language.
> This would allows an archive file containing many files to be downloaded at once, and be made available to the page.
>  
> This could reduce the http traffic for a typical web page to as little as 2 http requests.
> For https requests it would reduce the number of times, that the browser would need to check the content for being authentic.
>  
> I really don’t care what you call this tag, but here is an proposal:
> I am not limiting the archive format to be *.gz either, it could just as well be a tar, jar, zip file or even a lib
>  
> <archive src=”archive_of_css_and_js.gz” alt=”These files are static.”>
> <archive src=”archive_of_images.gz”        alt=”These files are dynamically changing often”>
>  
> These .gz files contains dozens of files, which would be placed into the browser cache just as a single download file would be.
> When a reference to a file is made, then the file would already be in cache.
> There is one other thing I should mention, normally when each file is fetched, the header attributes such as expires, cookies, etag, … are tied to the retrieval.  In the archive file, it should be possible to have a second file for each file which sets header fields tied to the member file.  This is bridging the gap between http and html.
>  
> When these files were needed they would already be in the browser cache.
> <script src=”abc.js”
> <img src=”abc.gif”
> <video src="abc.mp4"
>  
> Thanks,
> Peter Lenahan  Peter_Lenahan  at ibi dot com