- From: Edward Gerhold <edward.gerhold@googlemail.com>
- Date: Thu, 14 Apr 2011 22:23:11 +0200
Dear Working Group and Subscribers, i still think, it makes sense to upgrade the cache. Currently i am watching Coukumas Video about the cache again, i have the original document hixie writes in front of me as well as a pen, at the moment i have no text markers i use rather and often anywhere else, and i have in mind to write down the document for. When i opened the spec again, i thought "oh my god, how can i insert my idea here?". Then i remembered the content of the chapter, then i read the IDL definitions and the step by step algorithms and everything became clear again. Last mail i?d been wrong, i?ve read the inspector again. The Master, Manifest and Explicit entries lighted up as mime type in my mind, but only the cell?s name "Type" has been right. Considering the last chapter in the spec, Disk Space, i think "Yes, should be with quota." Well, what i want to say is, i read the doc, listen to the arguments in the video again and again, mark the places and spaces, learn the terms, try to write my document, i also though about directly editing ians document, i guess i?ll have to do both, i?ll write down the cases, i need the enhancements for, to give some reasons, i?ll write down the interfaces, and i?ll explain step by step what the algorithms do. Not like in my mail, i see it here in the original. This is easy, anyways, i can program the little bit i want added, so i can write down the algorithms step by step. Anyways, this is some work for me, too and i do a lot of other stuff, too, so i won?t post it tomorrow. But i STILL think, it makes sense. Hehe, in the video i heared, that a bug was fixed, that if whitelisted entries had a fallback, the fallback was always used. After that, he reminded me of the http:// prefix, this round i already forgot the words he said, but it reminds me of testing the http://linux-swt.de/index.php again in the NETWORK section, as, he said, the network prefixed urls never come into the cache. But - it didn?t work before. If this works, i?ll write this down for my users and say "Yo!". But it wouldn?t change my mind about feature completeness and giving us programmers full access to the cache. I could have skipped this argument in any case now, i didn?t want to downgrade, what i say. Well, i am working on the concept, to state it clear, what i (we) want to have in the cache, the current situation is good, it is usable for certain cases, but not very good, because the constraints are too large, and there is no programmatic access to the array with the manifested files. There is a need to grow and shrink the list, to add pages, to update existing and to remove pages (or media) from the list. As the mime isn?t saved, and i don?t want to think about sniff the content for each entry, saving the http request, i discarded again together with agreeing to a local identifier (a rename url function was my first thought, but a local identifier is right), again, saving the http request is meaningful, too. (Feature completeness). Well, i?ll write down the whole thing and post it. I don?t know how far i can edit the original document and strike the phrases through and insert the new. I am not so far today :-) But i?ll keep it in mind and on my list. Hehe. Ok, i?m gonna rewind the video (he is within the questions) and continue reading the original. With friendly regards Edward Gerhold
Received on Thursday, 14 April 2011 13:23:11 UTC