[whatwg] Enhancements to the Application Cache (Reminder)

Edward Gerhold edward.gerhold at googlemail.com
Sun Apr 10 14:04:14 PDT 2011

Hello Randy,

2011/4/2 Randy <randy at prowebdesign.nl>

> Seriously dude, you cannot send almost the same message for 5 times.
> Of course not, but nobody heard ever any word from me, and i didn´t know,
how far the list is read by the regular subscribers or the working group
itself. But it made sense, i´ve read on google (not in the blog), that i´d
been added to the blog, searching for the improvement of the cache.

> One time was enough and depending on the feedback, you could start talking
> about the subject.

Thank you. Below you give me an entry point, to try it again.

One other thing you need to figure out. If I have a master index.php that
> handles basically every page (like CMS's do), how would the dynamic caching
> handling those different states WITHOUT executing PHP (or ASPX, PERL, etc)
> on the client. The caching won't be capable of interpreting the serverside
> language.
What i want to cache is the page which was created be the content manager.

A programmer has to add some code to the CMS template or somewhere else,
that you can invoke the add function.
The page would be added to the cache.
If i cache the start page and open an article and cache it too, go offline,
and read the cached frontpage again, and click
on the link of the other page i cached, the url should be read from the

To give me the possibilty to grow or to shrink the cache, i should be able
to remove pages, too.
Modifying the page has an obvious sense.

The enumeration is needed for the control of the urls in the cache. Think of
writing a good application handling
the offline pages completely. The current situation is, that there are a few
HTML5 applications and everyone has
started to work with, with the impossibilty to use the cache like i need
today, the applications using it have to be
invented later, when the changes to the cache have been made.

::On a second read, i understood, that i´ve got you a little wrong. You mean
an index.php handling any page,
where the url is always index.php? Right. I´ll give it a try here.:::
Let´s guess, all parameters are passed over the post method. I can´t
separate data and url parameters giving the name,
e.g. if other data is coming with the request.
But writing the parameters with the url like index.php?post=1024&cat=512
should lead to the same page created.
I see, here i have to think again.
A compromise for rewriting urls could be a good solution. Think of writing a
web-app is programming work, so the code
for rewriting the urls has to be written completely by the programmer with
all the loc it has. The urls could be renamed and
replaced in the cached page. The only thing, is that this is using a few
more loc or a regexp and a replace for a quick solution.
But as long as the cache is only storing the few static files i can only
speak of it hypothetically. I know i come late with it in 2011
short before the last call. But better now than never.
::I´ll think about the index.php handling all.::

For the events i suggest two things. a) add/modify/remove and
begin/seek/pos/forwards/reverse/end fire events.
You get the data for or from the cache first.
Then you fire an event.
b) If the event data is different from the data you passed to the event, you
write that back instead of the data before.
That makes plugins for handling the appcache data possible. Programmers can
make a lot out of it. More than adding
If you don´t want to change the cache, take the event data, copy it to
another variable and modify this. The event data
will not be touched and not be written back, but the original you requested.

The last paragraph is for me and my mail.
I´m sorry, that i have not called for ten days and opened my sorted mailbox
to look for replys today the first time. My
document from last week is still in the old version, i planned it for today,
but had to do a lot of other things, that i opened
it an hour ago. Your email gives me a possibility to answer, before i place
the clean document again. I corrected the thing
with the writeback of event data, which i left a little bit unsure about on
the first of april. And i want to remove all the arguments
not needed and to beautify it, to be more readable. I am going to place that
soon again on the list, but like you all want, only once.
I have been registered, i noticed.

With friendly regards
Edward Gerhold

More information about the whatwg mailing list