21 Jan 2001 mattbradshaw   » (Journeyer)

i apologize in advance for the extremely lengthy entry... guess there's lots on my mind or something. maybe even some of it is a bit interesting...

i have to admit that i was tremendously amused by Iain 's desire to be stalked by a normal and attractive person. :)

dave winer describes his recent emancipation from cable tv. man, do i ever agree that tv is such an appealing waste of time. i can't imagine how much more productive i would be if it weren't for cable... need to do something about this! no more veg-ing! no more veg-ing!

in other news, i finally got my (wonderful) christmas present from my (wonderful) sister in the mail today. double cd of famous jazz performances and its pretty darned sweet. i forgot how much i enjoyed listening charles mingus. also included are miles davis, herbie hancock, t. monk... ahh, this is my soundtrack for the next year.

now that i've been (obsessively) reading a lot of the userland folk's weblogs, i came to the point yesterday where i really wanted to try out dave winer's outliner, radio userland, that a lot of em appear to be in love with. but i didn't (and don't) want to find some windows box to run it on (darned win32 and mac only sw!)... so i decide to try wine (again). wine seems to run quite a lot more (and more stably) than a year ago. but i was unable to get radio userland to work. :( so then i want to run yet another win32-only app, groove, and so i thought i'd roll up my sleeves and try to fix the relavent parts of wine. and that's what i've been trying to do this morning (and now afternoon, man time goes fast). well that and attend a short wedding. oh yeah, and listening to a bit of our new president, gwb. but i don't think i'm progressing that quickly at 'getting' the architecture of wine. if this works, i can promise that i'll have a dorky smile on my face all weekend. :) oh yeah, and this is why open source works if you already didn't know.

my wrists are starting to feel a little better as i've not had too much computer time lately, just a lot more design and reading stuff away from the wicked keyboard. which is fine. last sunday, while doing quite a bit of drumming i felt a little too much numbness for my liking. if it persists, i'll probably go see some doctor. or try acupuncture(sp?) or something.

i haven't forgotten my open payment protocol project (in the flavor of set), but i sure haven't posted much here about it. not much new to report really except that i think i've gotten a pretty good design with about 90% of it done. if you want to take a gander at it, give me an email or something. that should get me off my rump.

Matt's Musings... weird thoughts below


in tech related thoughts, i've been thinking about the (original) design of the web and pondering, "what might have i done differently?" i think the web has achieved such a level of acceptance as a network platform that many limitations present in its design and use are (often) unchallenged and ignored. let's challenge the assumptions, people! because it's fun and maybe something can come of it... oh yeah, and i apologize for my inarticulate ways. :)

my chief objection to the web's design (at present) is that it imposes an extremely one-way communications model. what i'm complaining about is that there exists NO mechanism by which a server can initiate data tranfer to a client. i still wish to maintain the traditional client/server model in which the browser initiates a request for a resource, but i do want the server to have the ability to send new data to the client to update the document. notice that this isn't push technology, as the client would only receive this new updating data from the server during the lifetime of the document being viewed. so instead of closing the connection when the document has been completely transfered, the client would leave the connection open, listening for any updates. the server could then send snippets of js specifying changes using the DOM. i think this would be really helpful, but it does muddy the waters of simplicity and elegance found in the traditional model of simply tranfering a 'document' . in particular, it places a quite heavy burden upon the server, maintaining state and all.

argh! man, when rereading the above in my diary preview that sure sounded crappy. pictures are needed, imo. gosh darnit, why can't standard web tools have a whiteboard! ... back to my rant.

if we look at the context/history of the development of the web, back in the day we weren't thinking in terms of web applications, or at least i don't think that we were. we were thinking in terms of simply transfering documents . so once the document was finished being transfered to the client, the connection and web transaction was done. at least until the client initiated another connection. and there's no ability for the server to send data any more data to the client. in order to make 'the network the computer', we'd have to rely on clumsy polling techniques. there's no way that the server can send an 'interrupt' to the client in other words.

these ideas, which have been floating around my head for some time, came to a point the other day when i was going to implement a web-based (because it's the universal network platform, remember! :) game of scrabble, my wife's fav. game of all time. sure, we could just do it (even easily) in java, but that is way too heavy for my liking. i think that web applications, really lightweight, using insanely familiar constructs are the way to go, or at least try. there was (and is) no way to let the server inform the players (using their handy web browsers) that it's their turn to try and figure out some words. argghh... i thought it might be fun to create an rpc server in moz that the web server could contact, but that is one ugly kludge if i've ever seen one. the above described mechanism of updating documents seemed to be the prettiest solution i could come up with.

i read a term on dave winer's blogs that i thought exceptional at describing the above functionality. "the two way web" but alas, his great meme candidate has another meaning. oh well.

of course, there is great value in 'just being a document viewer', and maybe these web apps should be called net apps and we should just use a network aware interface other than the ever-present web browser. next topic...

another unrelated objection to the present day web is its completely primitive user interface machinery (widgets). sure, i love the simplicity and the elegance that is the widget set found in html 1 (textbox, button, drop-down box, etc etc), but another set of standard widgets needs to exist. jon udell (sp?) mentioned this void a while back in his column at byte.com although some of the stuff he says about xul isn't accurate, imo. is anybody 'fixing' this? where's the w3?

i guess the above is more like an article draft (that's poor!) and not a diary. but oh well. please forgive! :)

anyways, i've rambled long enough. i find this refactoring of the web very interesting and invite any reader's comments. please let me know if i'm insane. thanks! off to the the dishes...

Latest blog entries     Older blog entries

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!