- From: SerpentMage (Christian Gross) <mailing@devspace.com>
- Date: Fri, 12 Sep 2003 13:18:54 +0200
- To: W3C Public Web Plugins List <public-web-plugins@w3.org>
Email has to be viewed in fixed width formating... I have been thinking about this patent for a while and what seems to be very unique about it is that a file is downloaded and it is associated with an executable like MIME. My idea is why not remove the MIME resolution and do everything in a different combination. My idea is to make the browser dumb and run something called an Internet Archive automatically. So anybody, have a read and maybe comment... (Just an idea..) We create something called a Internet Archive. The Internet Archive is a file that acts like an file system within a file system. The Internet Archive has a fixed length header, and body. Within the header are references to other files of the archive. EG it would look as follows: "File system" 1 2 3 4 5 [Header ][12345678901234567890123456789012345678901234567890] a=1 b=17 c=39 aaaaaaaaaaaaaaaabbbbbbbbbbbbbbbbbbbbbccccccccccccc The HTTP Client and HTTP Server are HTTP 1.1 protocol. The HTML page contains a reference as follows <html> <content ref="some_uri.ia" width="32" height="19" /> </html> When the client downloads the HTML page and hits the content tag the data is downloaded like an Image tag. The difference is the Internet Archive references a compressed set of files that the HTTP client knows about. This is similar to how a zip file. To ensure a certain level of security the file would be signed. What is special about the Internet Archive is that the first file in the archive is executable content like one would click on an icon on a desktop. Consider it like issueing an URL, not specifying the file desired and getting the default file (eg default.html). The client would download the content and then start the executable giving the executable a handle to the downloading(ed) archive. There is no resolution of mime types or anything along those lines, it is an automatic execution. Once the executable content has started and retrieves a handle to the Internet archive the running content would manipulate the archive data however it pleased. An optimization on the client side would be to cache Internet archives and the content that they reference. For example consider the following scenario. File: /archive1.ia [Header ][12345678901234567890123456789012345678901234567890] a=1 b=17 c=39 aaaaaaaaaaaaaaaabbbbbbbbbbbbbbbbbbbbbccccccccccccc File: /archive2.ia [Header ][12345678901234567890123456789012345678901234567890] a=1 D=17 c=39 aaaaaaaaaaaaaaaaDDDDDDDDDDDDDDDDDDDDDccccccccccccc There are two Internet Archives, which reference the files a, b, c, and D. Using HTTP 1.1 the header of the Internet Archive would be downloaded first. The HTTP client would inspect which files are in the cache and which are not. Then using HTTP 1.1 the individual files would be downloaded using byte addressing. EG Downloading both Internet Archives would be done as follows: ***** Http 1.1 Get /archive1.ia ranges-specifier = 1=1-32 ***** Client has downloaded the header and is inspecting what to get ***** Http 1.1 Get /archive1.ia ranges-specifier = 1=1-17 ***** Client executes first file, and gives a handle to the Internet Archive, which is processed by the running executable ***** Http 1.1 Get /archive1.ia ranges-specifier = 1=17-39 ***** Http 1.1 Get /archive1.ia ranges-specifier = 1=39-50 ***** Now user browses to different Web Page ***** Http 1.1 Get /archive1.ia ranges-specifier = 1=1-32 ***** Client has downloaded the header and is inspecting what to get ***** Client retrieves content from cache ***** Client executes first file, and gives a handle to the Internet Archive, which is processed by the running executable ***** Http 1.1 Get /archive1.ia ranges-specifier = 1=17-39 ***** Client retrieves content from cache ***** The identification of the files could be considered like cookies or URL references. An optimization could even to rename the Internet Archive with a new URL... Eg IA://archive1/identifier Where running the URL IA://archive1 would automatically download and execute the "default file" Christian Gross
Received on Friday, 12 September 2003 07:19:00 UTC