This is just a page that I publish things to when I'm bored, or put links to bits of information that I need to find again. A bit like a Weblog, but more random. Also: latest diff.
(Other) cool places to jot stuff down:-
With these places, why do we need Web servers at all?
The whole world loves a modal tuning. Modal tunings are basically suspended chords (when strummed open). Open tunings (Major Chords) are quite neat too. Yes, regular Spanish tuning is good because you can play in lots of different keys, but you don't get the same resonances as you do with open tunings. Imagine playing "The Rain Song" in standard tuning.
I have a slightly bizarre setup on my guitar... I tune it down an entire step (i.e., two semi-tones) from EADGBE, and then put a capo on the second fret. That's right - the capo is usually on the guitar, except when I have to tune down for something like "Yesterday". It means that there's less tension in the strings, and some purists will argue that it makes the guitar sound "flat" (in ambience, rather than pitch - it tunes just fine). However, it also reduces the action (makes the strings closer to the fretboard), brings the frets closer together, and generally confuses the crap out of anyone who tries to play on the guitar. So don't try to get me to change...
I don't tend to collect rare birds, but I do collect rare songs, such as Number One, Bob Dylan and the Band; an 8+ minute version of Heroes and Villains, the Beach Boys; and Say You're Gonna Leave Me, Led Zeppelin. I also have early versions of Strawberry Fields (as well as the ones on Anthology), and a minute or so of Carnival of Light, both (of course) by The Beatles.
The Yahoo! Games stuff is pretty good.
What would a perfect browser be like? (The rant that follows is modified from one on #swhack).
Browsers are just too cluttery these days. Bookmarks? Who cares - if we had better HyperText editors, bookmarks would be pretty obsolete. And this chaining of components... I just want a browser, not a browser/editor/email/irc/whatever combination.
My "perfect" browser would probably have just two modes - browse Web, edit text. Because I also think that text editors generally suck. Once again, most of the time I just want something that loads quickly, and can do simple stuff like search and replace. If I'm going to do anything complex, I'll run a script of some kind on it.
A WebBrowser/TextEditor combination would be great. That way, you could have a mode to show a resource as text/plain (kinda equivalent to view-source), and then edit the thing that you have in the browser window, saving the result. So it'd have to be able to save over HTTP (using PUT), and FTP, and locally. It would also be neat to have a redirect function, so that when you save to http://example.org/blargh, it actually saves it to the relevent FTP server. Of course, you'd have to beware of bloat, but something simple like that should be alright.
Also: utility-proxy.
Other random stuff... [SBPMISC]
$ u googlecount alestake alestake: 79 $ u googlecount sæstrengur sæstrengur: 33 $ u timenow 20020330-022130
#xpointer(string-range(//p,%22%5BSBPMISC%5D%22))".If you care about the ordering of Þ, the diff. 'twixt em and en, nouning and so on, then good for you.
Although I'm usually quite keen on preserving anachronistic spelling/ grammar/punctuation conventions, I think that the use of hyphen-like characters can be deduced unambiguously from context, such that it renders em/en/etc. rather obsoltete. Using a double hyphen for an em-dash seems rather petty to me when an ordinary hyphen-thingy with a single whitespace character either-side will be interepreted just the same.
Asterisks should almost certainly be called splats. That got me wondering about the semi-mythical Stanford extended ASCII circle-plus character, too.
Phenomic is, of course, also phenomic.
Python is currently my favourite programming language [period].
But I do feel as if I'm letting Larry down by not getting into Perl... I mean, the guy comes up with so many great quotes. Anyway, I was delighted to learn that now Google has given us the Usenet archive (thanks, Google!), we can search out all of the old Wall-ism originals.
O.K., so perhaps STAR is my favourite programming language.
The GAWK Manual is an interesting read, for the quotes, and the introduction (to me, when I read it) of the term "dark corners".
Mixing languages is fun; a ToII, and evolution. cf. schema-hacking.
But what's a language?
[...] a a namespaces is a language: a set of names plus a set of syntactic constraints plus - to be useful - a meaning shared by writer and recipient.
That meaning can be to a certian (and, with technology, growing) extend by expressed formally in a mapping to other languages. Or is can be disccussed in the bar, a mid: URI written on the back of napkin, and used from then on by the people who were in the bar in the messages they write to each other.
- TimBL
Ruby:
RDF profile for XHTML stuff: what metadata is required? author, authorName, authorEmail, authorHomePage, lastModified, created, alternate, mirror (alsoAt), copyright, version, previous, diff, help, meta, index, partOf, contains, checkMarkup, bookmark, toc, help, glossary, home, derivedFrom, obsoletes, search, toc, persistencePolicy [...] cf. Link types, by Fantasai.
<!DOCTYPE: necessary or not?
Sometimes you really do want to extend HTML. There are just so many annoying little imperfections in HTML, it's really difficult to avoid extending it - sometimes even inadvertantly. Of course, this leads us into discussions about exactly what consitutes a language, how the various syntactic forms of XHTML relate to the namespace, and language mixing/evolution etc., but there should be a short list of practical deployment tips for XHTML. XHTML Modularization is supposedly there for one, but it's difficult to set up validation because XHTML is such a complex beast. Perhaps the law of free extension under namepaces is actually not that bad? I've resisted it in the past, and I'm sure to resist it again at points in the future, but sometimes it does seems to be the best option. Of course, Sands (Syntax and Semantics in XAG) tries to ground the logical structures behind documentation schemata in the Web, but I doubt that it does all that good a job of it.
Idea: a little SVG logo "[SBP]",
and then a normal image nested inside an <object> element.
Uses XLink!
There are some classes that I use fairly regularly: "toc", "submit", "navbar", "k", "note"... it would probably behoove me to document their general usage, and come up with some stylesheets.
I could also create a "class" Class in RDF, to describe the NMTOKENs... I wonder if that would relate to Sandro's uname property proposal (cf. discussion).
ul.inline { display: block; margin: 0em 1em; padding: 0em; }
ul.inline li { display: inline; }
ul.inline li.li:after { content: "; "; }
DanC's excerpting page is interesting, because excerpting is part of the Web. I also find it interesting that the page has no <h1> heading, and is being used as an RDF namespace. cf. my little finding on www-archive.
My clipboard is a rather important application for me. I also use IRC channels as a bit of a clipboard (mainly #sbp).
I just worked out something about quoting conventions in plain text. When people give me quotes like the following:-
[[[ "Once a long time ago [Jean Louis-Gasee] asked me in his French way if I was a pimp or a whore. I had trouble answering, but when I turned the question back at him, without hesitation he said he was a pimp." ]]] - Aaron Swartz, quoting Dave Winer
they don't make sense to me, because once you remove the [] section, it doesn't read correctly. Now, obviously what's going on here is that the [] section is a replacement for a missing word; in this case "he". Anyway, it got me thinking about my quoting style, and so I thought I'd write up some general rules.
Quotes with newlines should be [[[ triple square bracket quoted. ]]] - and carry a citation.
Well, that's it. Thanks for reading. For more information, consult your local style guide.
Using notepad. When it gets bigger, I'll probably use Wordpad. Amaya is no good, and VIM is a tad bitty.
Generated from nothing at the moment. RDF is an option, but it's really a data format. One advantage of generating is that I'd be able to save different versions as I go on, but then again, I'm not too interested in capturing every little word and idiom. This is a representation of the resource, and the resource is all that matters.
Variants of HTML: perhaps XHTML 2.0 will be good, or perhaps it can be modified to be more like XNote, or perhaps bits of it can be incorporated in a new XHTML 1.x variant. Or perhaps namespace pure XHTML is the way to go.
"New" format for expressing the current time, for use in identifiers: timenow. YYYYMMDD-HHMMSS. Python script to get it:-
#!/usr/bin/python
"""Gives the time as a 15-char string"""
import time
def timenow():
return time.strftime('%Y%m%d-%H%M%S', time.gmtime(time.time()))
if __name__=="__main__":
print timenow()
I have an alias set up to run it as u timenow - prefixed by
"u" as all of my utility scripts are. Then, to archive a file: cp
blargh blargh`u timenow`.
TODO: Define a new MTRUID (More Than Reasonably Unique ID) using timenow and the MD5 hash of some
random information. Something like timenow+md5(timenow+rand).
I want to get a mid: script running so that I can put an mid into my browser, and have it return the message, if it can find it. Places that one may find it include:-
If all of these fail, a simple Google search for the "mid:" itself would have to suffice.
The Plex is an information space idea, based on P2P, URI-tuples, and hash functions. You'd better get used to addressing files by SHA 512 (5**); it's a decentralized SHA => file service. I found a good program for Windows that will get the hash of a file.
$ sha512 abc SHA512 : abc ddaf35a193617abacc417349ae20413112e6fa4e89a97ea20a9eeee64b55d39a 2192992a274fc1a836ba3c23a3feebbd454d4423643ce80e2a9ac94fa54ca49f
Good document structures under Windows, like under *nix, and under OS X.
Miscellaneous \Mis`cel*la"ne*ous\, a. [L. miscellaneus mixed,
miscellaneous, fr. miscellus mixed, fr. miscere to mix. See
Mix, and cf. Miscellany.]
Mixed; mingled; consisting of several things; of diverse
sorts; promiscuous; heterogeneous; as, a miscellaneous
collection. ``A miscellaneous rabble.'' --Milton. --
Mis`cel*la"ne*ous*ly, adv. -- Mis`cel*la"ne*ous*ness, n.
- Webster's Revised Unabridged Dictionary (1913)
What's the deal with permalinks? They're IDs in HTML WebLogs that are meant to be persistent, so that people can link to them. Using a little symbol for that is traditional.
Why validate? Well, sometimes I like to strip junk out of XHTML using XSLT, and at other times, I like to think that odd browsers like Amaya will render this page correctly. Of course, XHTML and FPIs hork, but that doesn't stop me from creating my own little DTD and popping it on the Web.
Some people may resent the fact that we have a wonderful HyperText Web, and yet I'm feeding all this information into one big document. Isn't that bad? Well, usually I'd agree, but the fact is: publishing to the Web is not easy. Maintaining lots of pages and making sure that all the links work is quite time-consuming, and so often, when I come up with something, I don't bother to create an entirely new page to express that thought, and it gets wasted. At least with this huge page, I just scroll down to a place that I'm familiar with (or even an arbitrary location), and clean it up at a later date, if at all. It gives some freedom.
Of course, that approach is bad because it means that a) people won't be able to find information as easily, and b) it means that the file gets very large, very quickly. Well, this file is 80% for my own use anyway, so I want it to be searchable... and it is. You can convert it to text only (using the W3C/Aaron's online service), and then grep it.
def stripmarkup(s): tags = re.compile(r'(<[^!][^>]*>)', re.S).findall(s) for tag in tags: s = string.replace(s, tag, '') return s
As for the file size... well, I don't particularly mind if it gets as large as 50-100KB. People should be able to cope with that, going by today's average connection speed figures. I'm not sure what I'll do when I need to prune it - probably just move it into another archived file. Perhaps even spew it at www-archive.
Documents should have statuses (cf. Give the status).
The status of this thingy is that it's an XHTML (text/html) representation of some resource on the Web (more specifically, a Generic Document), as denoted by the URI http://infomesh.net/misc
The semantics of that resource are not all that specifically defined. This page is mainly for my amusement only, but if some others get a kick out of it, then that's nice too. Don't expect the content to be stable - it's whimsy. If you want to link to a stable version, use Alexa (if you trust it not to go down).
