mircea_popescu: http://btcbase.org/log/2018-02-03#1780757 << all they're good for. ☝︎

mircea_popescu: mod6 what of ?

mircea_popescu: ben_vulpes hey, imagine all the idle posturing going on in every tiny social group. kids used to be hackers, after being secret agents, after being you know.

mircea_popescu: in other lulz, the day's log page is fulla non-linking links as a result of http being the discussion

mircea_popescu: aaand thereby log eaten. mazel tov.

mircea_popescu: http://btcbase.org/log/2018-02-02#1779704 << i recall looking into it few years ago and not being impresse.d ☝︎

mircea_popescu: you'll also probably have trouble without a connection: close signal, and so on.

mircea_popescu: ://trilema.com/2014/ill-pay-for-your-tits/ as a complement to 302 response also quite useful.

mircea_popescu: http://btcbase.org/log/2018-02-02#1779676 << this is not so. looking through trilema's : X-Pingback: http://trilema.com/xmlrpc.php is essential for the whole pingback functionality, which i deem a prime class. off top of head cache-control is probably also useful, as it tells the client the very useful bit of whether server deems resopurce is type 1 or type 2 in http://btcbase.org/log/2018-02-03#1780528 sense ; Location: http ☝︎☝︎

mircea_popescu: http://btcbase.org/log/2018-02-02#1779674 << not practically. 2kb max size or what was it, won't get you far. ☝︎

mircea_popescu: but you don;t know what to put in the where to begin with. so you get potentially boundless piles of metadata.

mircea_popescu: right, everything is selects in the end,

mircea_popescu: yes, but the thing still starts with describe or w/e it's called

mircea_popescu: but anyway -- you can even now make a server return ~infinite headers if you wish. ☟︎

mircea_popescu: (practically, they just didn't feel like making one verb per thing, because it'd have ended all sgml.)

mircea_popescu: well you don;t know what you don't know. exactly as in the slavery threads.

mircea_popescu: trinque because you don't know what questions may be asked, is the idea.

mircea_popescu: through header too.

mircea_popescu: this is a major flaw of the "design" such as it is, that yes you're asking for potentially infinite metadata.

mircea_popescu: similar to trinque goes into santaria shop, "do you have any bibles ?" "yes, shall i read you from it ?" "no, thanks. how much for one ?"

mircea_popescu: and if you say "i'm interested in x but in 0 bits of it" the logical implication is that you're only interested in any meta-x may be available. which yes might be null, but also doesn't have to be.

mircea_popescu: you're asking for 0 bits of X, not for 0 bits altogether.

mircea_popescu: but that's not the meaning of the 0.

mircea_popescu: the reason i linked the two items was, that you had a paradoxical result evaluating logical implication of negated absence in there ; and a paradoxical result evaluating predicated nothingness, here. seem to me intuitively they're related to some sort of 0-difficulty you have.

mircea_popescu: i'm lost ; say what ?

mircea_popescu: hm ?

mircea_popescu: "our data type is vector list array, tyvm & come again!"

mircea_popescu: fancy that wonder, text, html and utf-8. some data type.

mircea_popescu: trinque the most reliably encountered header in responses is Content-Type: text/html; charset=UTF-8

mircea_popescu: (good thing we got the cloud collaborate suite from $corp, too... how could we have gotten this far without ???)

mircea_popescu: well the whole thing is one huge napkin doodle shared over the chan as it stands ☟︎

mircea_popescu: "don't have a head, use get size 0"

mircea_popescu: basically whole thing was "Don't implement a special show table status, use the underlying select item from schema directly"

mircea_popescu: right. which is what was contemplated here too.

mircea_popescu: and this show table status is a proxy for select autoincrtement from inf schema

mircea_popescu: that one is select john from tabler and the other is show table status.

mircea_popescu: isn't this "tell me about this table for a total of 0 rows" ?

mircea_popescu: for the autoincrement metaconstruct.

mircea_popescu: but i can get, eg, a table's current index position

mircea_popescu: so sql doesn't have an equivalent of head is the idea ? or doesn't have an equivalent of get 0 size ?

mircea_popescu: explain this to me ?

mircea_popescu: yeah, it's not clear that the extant stack of shit is capable of supporting some sort of meaningful practical consideration.

mircea_popescu: trinque i meant strictly from a logical perspective.

mircea_popescu: http://btcbase.org/log/2018-02-02#1779651 << the correct statement would be, "you have experiences in good faith and i have imaginations in good faith, that i don't bother to check nor do i ever sit down to reason as to what THAT implies re faith" ☝︎

mircea_popescu: and considering what happens in response to curl -v -X GET -H "range: bytes=1-8" www.loper-os.org (full page dump) we can add asciilifeform / nfs to the list of "does not support ranges" ☟︎

mircea_popescu: for instance here's what your server says : http://p.bvulpes.com/pastes/XGPpg/?raw=true

mircea_popescu: http://btcbase.org/log/2018-02-02#1779643 << trivial to check, issue a head, see if "accept-ranges" is in there. if not -- no support ☝︎

mircea_popescu: so yes, head support universal, ranged request support present but not that reliable.

mircea_popescu: http://btcbase.org/log/2018-02-02#1779633 << all servers had head because it's antique ; ranged requests went through an ebb and flow of implemented, then disabled because implemented wrongly and buffer overflows, then etc. ☝︎

mircea_popescu: http://btcbase.org/log/2018-02-02#1779620 << is this a rehash of >> http://btcbase.org/log/2016-12-16#1583902 ?! ☝︎☝︎

mircea_popescu: if i tell you i wish to see the file x except not any of its content this is logically equivalent, quite directly, to "metadata only plox"

mircea_popescu: http://btcbase.org/log/2018-02-02#1779611 << how do you reason ? ☝︎

mircea_popescu: as that deed was stepping towards, yes

mircea_popescu: we still need js as much as last time this was discussed, a week ago

mircea_popescu: quite opposite of sense ; and a much more important ablation target than forgotten/disused http words.

mircea_popescu: or at the very least so "page is not usable until our shit is in"

mircea_popescu: asciilifeform oya.

mircea_popescu: if my loading of a web page dies 10kb in, i want that 10kb to have contained the article, not some "template" fugly.

mircea_popescu: not evident that this should be as it is.

mircea_popescu: currently there's no equivalent hierarchy of html documents, so they are always sent in order of importance of parts.

mircea_popescu: (most important first etc)

mircea_popescu routinely pulls data from slaves/processes/etcetera with an explicit or implicit size limit and strong conventions as to ordering

mircea_popescu: http://btcbase.org/log/2018-02-02#1779605 << there is a good argument in favour of overloading meaning into verbs rather than inventing new verbs. an implementation of head as "get with 0 bytes accepted" is actually an improvement over current. ☝︎

mircea_popescu: anyway, this worth moar thinking.

mircea_popescu: and the old handles will 404

mircea_popescu: "immutable" still enforced by server only. tomorrow i can decide to make a whole diff tree and gl to you.

mircea_popescu: tomorrow, linkrot will thereby mean "you might not have the original data ; OR you might not have the original program".

mircea_popescu: not so. consider : linkrot today means, "you might not have the data original server had". like you lost lispmachine FET say.

mircea_popescu: it's not clear it'd do anything for linkrot sadly.

mircea_popescu: no longer a hodge-podge of "plugins" and "themes" haphazard and therefore entirely flat conceoptually.

mircea_popescu: consider the machinery that blogs become, in this system

mircea_popescu: it is threateningly complex.

mircea_popescu: so in point of fact saying "why would alternate views exist" is facetious, i listed like 4 in five minutes.

mircea_popescu: rather than unmaintainably clog out my blog and risk being deleted any day now.

mircea_popescu: now, this ad hoc hack could have been a proper alt-leaf thing

mircea_popescu: http://btcbase.org/log/2018-01-13#1770183 << this ☝︎

mircea_popescu: consider the hack i did recently for hanbot

mircea_popescu: i dunno.

mircea_popescu: "i want trilema as rss not trilema as main page" or "i want trilema as list of archives not as main page"

mircea_popescu: in fact, another ready example : rss.

mircea_popescu: i dunno, i confess i've not prior considered the v-http server.

mircea_popescu: possibly.

mircea_popescu: and it's a wonder if it'd be actually worth it. consider -- what alternative wp-mps would i maintain on trilema ?

mircea_popescu: not in loc lines necessarily, but in what lifting it actually does, conceptually.

mircea_popescu: this is by now SO MUCH HEAVIER than current http...

mircea_popescu: not mere "merkle tree", but actual proper v struct, with seals resources and errything.

mircea_popescu: but you WANT header then ; it is the evident place to report the v-struct

mircea_popescu: yes but this is a much more integrated v-http server than previously even hinted at.

mircea_popescu: "process my data with the press X as identified from your manifest, i don';t trust X'"

mircea_popescu: then you could in principle indicate WHICH LEAF you even permit.

mircea_popescu: only if http server runs v native.

mircea_popescu: seems to me it's a thing unto itself.

mircea_popescu: can you actually say the above snippet is "really kludge to display bouquets of statics properly" ?

mircea_popescu: the latter part is what we have issue with.

mircea_popescu: yes but...

mircea_popescu: "find a thing such as x" vs "follow steps x give whatever comes out" ☟︎

mircea_popescu: asciilifeform in that one is imperative and the other descriptive so to speak.

mircea_popescu: (for convenience, in pesudocode : select from an array of possible headers the one which corresponds to the remainder of dividing by 42 the number you obtain by converting from hexadecimal format the string you obtain by taking 6 characters from the 7th of the md5 hash of the current date)

mircea_popescu: it evidently crosses all conceivable sanity boundries. so ?

mircea_popescu: that's the major challenge : can the argument be constructed to see what is actually wrong with that piece of gnarl ?