a111: Logged on 2019-05-03 22:19 asciilifeform: Mocky: iirc mircea_popescu's scheme centers around a return to the good ol' days when ~all~ names could be found on a given machine's /etc/hosts (or equiv.) -- so one ~could~ in principle use a very slow mechanism to sync'em, and it'll still work in practice. ( tho i also would rather avoid cementing tcpisms into future design of whatever type )
a111: Logged on 2019-05-03 22:21 asciilifeform: Mocky: re vpatches -- as i understand, the contemplated structure is a linear chain, and 'siblings' could only occur in event of schisms/civil war , rather than as a regular thing
a111: Logged on 2019-05-03 12:19 mp_en_viaje: moreover, the selection of seals you put in your seals dir permits you to choose from a multitude of different trees.
Mocky:
http://btcbase.org/log/2019-05-03#1910995 << no. The proposed spec results in a linear chain, the tip of which is the only place where the next vpatch could go without a schism/civil war. On your own computer you can press wherever you want in the chain (or have private branches w/e), but in public you can only release further vpaches based on the same tip of the chain that everyone has. this is defacto consensus.
☝︎ Mocky: Maybe my prior comments were too abrasive / snarky making it difficult to engage, if so that's on me. Or maybe I just don't understand. Either way there's no point me trying to continue with the content of an argument that no one can engage with.
mp_en_viaje: Mocky, your ideas of what means "public" an' "private" are unsuitable in two fundamental ways.
mp_en_viaje: the first is that there does not exist a "public" or a "private" as such, in the manner there exists "good guys" or "the rootless cosmopolitans" as found in novels
badly written by tendentious authors.
mp_en_viaje: there's such a thing as the private1 of mp alone, and the private #2 of mp and slut x, and private #3 of mp and slut y, and private #4 of mp and slut x and slut y, and so following.
mp_en_viaje: the second is that however vectorial or otherwise non-scalar / non-naive extension of bad literature these "public" or "private" be defined, they are NOT properties of the user, deemed as an individual, sovereign and central item in the construction of the world.
mp_en_viaje: the novel does not follow hero ilya muromets (whom you know is the hero cuz that's who the novel follows) and from this perspective describes his adventures in the lands of public or private.
mp_en_viaje: rather, the novel follows ~the hero public~, and from THIS perspective tells the adventures of various cioloveks, ~whose names are only in the first place predicated on THEIR appartenence to the hero of our novel, the status-of-being-public~.
mp_en_viaje: for which reason : yes there is going to be a press for anyone pressing. how exactly this press is constructed out of the universe of possible pressings, that's a matter established through the interplay of domains, as defined by the keys belonging to them.
mp_en_viaje: (in particular
http://btcbase.org/log/2019-05-03#1911102 is at best naive : if you insist that my key does not sign competing chains [which is not even necessarily a bad idea] i will simply make ~multiple keys~, which will each sign whichver chain i wish to sign for whatever reason, and which will pass the "omg civil war" safeguard while achieving exactly no satisfaction of the bijection contemplated -- they'll still be my keys, one and all.)
☝︎☟︎ a111: Logged on 2019-05-03 22:21 asciilifeform: Mocky: re vpatches -- as i understand, the contemplated structure is a linear chain, and 'siblings' could only occur in event of schisms/civil war , rather than as a regular thing
mp_en_viaje: as it stands, i also do not see a problem with the same one key signing whatever count of ~compatible~ versions of the universe, where the key's to define what compatible actually means, following therefrom that there's exactly jack shit one can systematically do by looking at the totalized pile of keys signatures and possibilities. (the familiar will readily recognize a pattern here).
mp_en_viaje: now then, specifically as to the whole "no. bla bla", let's look at how this works in practice :
mp_en_viaje: At t0, there is a genesis, consisting of a gns file, containing a0 -> b0 ; and a manifest file, containing #t0.
mp_en_viaje: At t1, there is a patch upon t0#t0, transforming gns to a0 -> b0 ; a1 -> b2 ; manifest becomes #t1. this is signed by A, W, F, G, U, K.
mp_en_viaje: At t2, there is a patch upon t0#t0, transforming gns to a1->b1 ; manifest becomes #t1. this is signed by A, Q, G, B, N, I.
mp_en_viaje: At t3, a fellow looks to add a2 -> b5 to the gns. suppose this fellow whose name is R knows only Q. Thus therefore he presses t2#t1, and so gns becomes a1->b1, a2->b5 while manifest becomes #t3.
mp_en_viaje: suppose instead the fellow only knew K. he'd have pressed t1#t1, resulting in a gns a0->b0 ; a1 -> b2 ; a2 -> b5 and a manifest #t3.
mp_en_viaje: suppose yet again the fellow knows A. in this case, he has a choice : either he does the first suppose above, or the second. this choice is ~his~.
mp_en_viaje: at t4, another fellow named F looks to change gns so that a0 -> b8. this fellow goes through all the keys, loads up all possible tips (which are t1#t1, t2#t1, t3#t3, the last one ~in itself ambiguous at this point in our tale) and produces : on the first chain a t4#t2 ; on the 2nd chain a t4#t2 ; on the third chain a t4#t4, and signs all of these.
mp_en_viaje: at t5, yet another fellow has the choice of whether to follow F's chain, or R's chain, or the A-Q-G-B-N-I chain or so following.
mp_en_viaje: where's the civil war ? obviously if A wants to be upsed at F, or R, or for that matter Q or U it's entirely his right and priviledge to proceed ; but what ~systematic~ rule for warmaking can you propose on the basis of this flow of history ?
☟︎ mp_en_viaje: (pro tip : of fucking course you can't, if you could it'd be therefore cryptographically weak.)
mp_en_viaje: to put it in other terms -- if this fit in head, it'd be broken. the ~mechanism~ is supposed to be fit in head ; not the data.
mp_en_viaje: speaking of which, at some point a while ago hannah wakes up in a panic -- she had a nightmare.
mp_en_viaje: the nightmare was that i had ordered her to document all the water.
mp_en_viaje: nobody's about to order anyone to document all the water -- and by this i mean, asciilifeform ain't gonna send himself on such foolish errands, nor you, nor anyone else. let the water be.
mp_en_viaje: sadly belgrade's sour milks and bread specialties culture seems to have muchly decayed in the intervening ~decade.
mp_en_viaje: this is a horrible loss to human culture i guess few are apt to appreciate. but horrible nevertheless.
mp_en_viaje: basically the cheap and inedible "fornetti" crap took over.
diana_coman: heh, fornetti is the universal name for industrialised replacement of replacement in the pastry world; I hadn't realised it had spread that far.
mp_en_viaje: poor serbs. they really had the world's best pastries last decade.
diana_coman: well, fornetti and nutella and whatever else, when going cheaper and then cheaper and so on, how else can it really go
diana_coman: hotelmoskva has no better internets than the public house apparently
mp_en_viaje: now, shockingly, it's sweetened & flavoured hydrogenated oils
mp_en_viaje: diana_coman, i guess maybe i'm being ddos'd ? nfi, by now kinda weird.
mp_en_viaje: in other sads, apparently yest trilema header was blank because i used - instead of _ in filename. pshaw.
Mocky:
http://btcbase.org/log/2019-05-04#1911138 << I did not construe asciilifeform's language of civil war/schism to suggest that the proposal is weak to systematic attack, merely that a schism between people could be reflected in the graph
☝︎ a111: Logged on 2019-05-04 05:44 mp_en_viaje: where's the civil war ? obviously if A wants to be upsed at F, or R, or for that matter Q or U it's entirely his right and priviledge to proceed ; but what ~systematic~ rule for warmaking can you propose on the basis of this flow of history ?
Mocky: mp, I can make my point using your flow if you allow a slight change to the patch at t1, such that instead of 'a0 -> b0, a1 -> b2' it does 'a0 -> b9, a1 -> b2'.
Mocky: supposing that a0, a1 etc. are host names and b0 etc are ip addresses, and further suppose that at time t2 a1 becomes unreachable at b2 and reachable at b1 (precipitating the patch a t2) then there are now no nodes in the flow going all the way out to t5 and beyond where both a0 and a1 would be reachable if pressed
Mocky: my point is not that this is likely to happen, but that it would defeat the purpose of host name resolution to release patches in this way and so it would end up being a linear chain among cooperating folks. But on further consideration I see that this is only true when patches contain more than mapping change
Mocky: "more than one mapping change"
☟︎ a111: Logged on 2019-05-04 05:31 mp_en_viaje: (in particular
http://btcbase.org/log/2019-05-03#1911102 is at best naive : if you insist that my key does not sign competing chains [which is not even necessarily a bad idea] i will simply make ~multiple keys~, which will each sign whichver chain i wish to sign for whatever reason, and which will pass the "omg civil war" safeguard while achieving exactly no satisfaction of the bijection contemplated -- they'll still be my keys, one and all.)
a111: Logged on 2019-05-04 13:38 Mocky: "more than one mapping change"
a111: Logged on 2019-05-04 14:31 asciilifeform: promises to be gnarly.
mp_en_viaje: consider education : you could sleep in class instead of paying attention. "but why would i ?" "you fucking wouldn't, obviously. but if someone did -- will you guard them ?"
mp_en_viaje: Mocky, moreover, there's no "one true internet" understood as a collection of equal and idempotent sites. maybe a1 no longer resolving is a feature not a bug for a subset of pressers.
☟︎ mp_en_viaje: (gns is gns, a's can be domain names as they can be dictionary words. i can see the argument for "global warming" or w/e tv "celebrity" not resolving to anything.)
mp_en_viaje: heck, half my discussions of such start with "that idiot one, with the horse face" or equivalent. i've been living for many years with tons of entries in the pantsuit namespace not resolving to anything whatsoever, and to no detriment.
☟︎ a111: Logged on 2019-05-04 17:54 mp_en_viaje: Mocky, moreover, there's no "one true internet" understood as a collection of equal and idempotent sites. maybe a1 no longer resolving is a feature not a bug for a subset of pressers.
a111: Logged on 2019-05-04 17:56 mp_en_viaje: heck, half my discussions of such start with "that idiot one, with the horse face" or equivalent. i've been living for many years with tons of entries in the pantsuit namespace not resolving to anything whatsoever, and to no detriment.