92.06.11 RC, TBL

Summary of HTTP 0.9

The current version of HTTP can be summed up as follows:

• A browser only sends the command GET followed by a server/document identification optionally followed by a question mark and a list of keywords (with no spaces). This we call a UDI
• A server replies to a GET by supplying a piece of ASCII text marked up in HTML, whereby just plain text with no tags at all is acceptable. [but should be prefixed with PLAINTEXT because the default style for HTML is (a mistake??) free format.]
• the server holds no state w.r.t. the browser.
• if a document contains anywhere the ISINDEX tag then the browser takes this to mean that a valid UDI from this document to another one may be generated by taking the current identifier and appending the keywords given in the search panel.

Note that if responses from the server are encoded into HTML tags, then HTTP is very asymmetric: the command GET is not in HTML. [Yes, but this is part of HTTP not HTML as you point out - TBL]

Advantages

There is nothing other than HTML marked up text coming out of a server

The browser does not know whether a server serves an index or a set of hypertext documents.

Disadvantages

There is no lexical distinction between what constitutes HTTP information, document contents, document kind.

In particular, error messages are returned as documents and may or may not be ignored by the user.

Good Characteristics

Both browsers and servers silently ignore things they cannot handle:

• the browser ignores unknown tags,
• the server ignores anything after the single GET argument (document/serverid?keywords).

This should be carefully kept: it means the user is not annoyed by messages he should not have to deal with (I am always thinking of the innocent, naive user who just happens to use WWW, not somebody hacking unix programs in C)