tmsr - Log 2020-03-12

trinque: if you want "builds crystalspace" to be the purpose towards which we're reaching, might as well install ubuntu.

trinque: if you want "doesn't have batshit bashism" to be the measure of which shell to use, idem

trinque: ok, so what rankles my ass about your goodness list is you have items at wildly different levels of the tree all flattened together like they're of the same importance and same level of the ontology.

trinque: this is a wall of hubris, my friend, and sorta flies in the face of what I was trying to convey with the series.

trinque: the "big dream" approach does not work, which is why I started simply, and proceeded outward.

trinque: but anyway. I wasn't describing an item I wish I had. I was describing an item I'm building.

trinque: I propose you delete that list, and you write another. In the new list, write down only the items that will get the machine to boot, and allow the user to edit and rebuild the OS.

ossabot: Logged on 2020-03-11 23:57:29 trinque: if you want "builds crystalspace" to be the purpose towards which we're reaching, might as well install ubuntu.

dorion: http://logs.ossasepia.com/log/trilema/2020-03-11#1959471 - my measure is reaches the point of supporting the implicit clients within a reasonable time frame based on where the clients are presently.

ossabot: Logged on 2020-03-11 23:58:37 trinque: for the record, my measure is "self-hosts"

ossabot: Logged on 2020-03-11 23:58:52 trinque: it can be used to build and improve itself. the end.

dorion: http://logs.ossasepia.com/log/trilema/2020-03-12#1959473 - I'll be the first to admit the tree can improve and my ability to improve it can improve. I erred on publishing and being clear it was a draft cause it had been delayed long enough.

ossabot: Logged on 2020-03-12 00:13:46 trinque: ok, so what rankles my ass about your goodness list is you have items at wildly different levels of the tree all flattened together like they're of the same importance and same level of the ontology.

dorion: http://logs.ossasepia.com/log/trilema/2020-03-12#1959474 - the first reason is the current gbw-node is in python. the second reason is t

dorion: he only genesis'd irc client, yrc is in python. jfw will be the first to tell you he'd rather neither were in python, nevertheless they

ossabot: Logged on 2020-03-12 00:13:55 trinque: what's python even doing there?

dorion: sorry for the busted lines. mp_en_viaje, diana_coman, does eulora use python for anything ?

dorion: http://logs.ossasepia.com/log/trilema/2020-03-12#1959478 - gales linux covers that today. I don't doubt it can be slimmed substantially. nevertheless, the take away from the first thread was it ain't enough.

ossabot: Logged on 2020-03-12 00:37:25 trinque: I propose you delete that list, and you write another. In the new list, write down only the items that will get the machine to boot, and allow the user to edit and rebuild the OS.

ossabot: Logged on 2019-11-12 16:59:56 mircea_popescu: so it's basically a training tool, as far as that goes, a didactic example

dorion: http://logs.ossasepia.com/log/trilema/2020-03-12#1959477 - what do you expect to run on what you're building ? trb ? or is it a bridge too far since isn't required to boot, edit and rebuild ?

ossabot: Logged on 2020-03-12 00:19:33 trinque: but anyway. I wasn't describing an item I wish I had. I was describing an item I'm building.

dorion: trinque ftr I very much appreciate what you're conveying in your series which is why I want to both talk about and see it continue. I'm not ignoring it, but I'm also not ignoring the mandate to support the implicit clients. how much are you taking the latter into consideration in what you're building ?

feedbot: http://trilema.com/2020/closure/ << Trilema -- Closure.

dorion: http://logs.ossasepia.com/log/ossasepia/2020-03-12#1020639 - I wrote that in #o cause I thought, "voice will be lost in here any moment now." but then I thought, "maybe that's the immaturity you have to kill to begin with."

ossabot: (ossasepia) 2020-03-12 dorion: just gave the latest article a second read. seems like he's saying, "you all could be men, but for whatever reason you're not and I've had enough of the retardation to interact with it further. perhaps me walking away is what's needed to wake you up."

jfw: http://logs.ossasepia.com/log/trilema/2020-03-10#1959363 - I *thought* there was something deeper being said in this thread that I wasn't grasping.

ossabot: Logged on 2020-03-10 21:26:33 mp_en_viaje: this is of course my problem ; you can do it elsewhere cheaper / better / whatever, i'm the last dood to get in the way of any such a thing.

ossabot: (ossasepia) 2020-02-18 diana_coman: http://logs.ossasepia.com/log/ossasepia/2020-02-18#1018737 - traditionally that is most obviously found out in crisis situations really; did you read modernism and traditionalism?

ossabot: Logged on 2020-03-12 01:31:41 dorion: sorry for the busted lines. mp_en_viaje, diana_coman, does eulora use python for anything ?

dorion: diana_coman ok. ty, I asked because it was mentioned in your notes, albeit with skepticism.

diana_coman: dorion: ah, initially it was I guess because of blender; but meanwhile blender will not get to do much/be supported anyway.

jfw: I haven't got an intro article out yet (sorry diana_coman) but genesis for the parts are at http://fixpoint.welshcomputing.com/v/ : gscm, gbw-signer, gbw-node.

mp_en_viaje: jfw, http://logs.ossasepia.com/log/trilema/2020-03-07#1959149 << was this more a "can i assume you have a complete republican computer, as per the [not yet written, let alone become part of common culutre] recipe" ?

ossabot: Logged on 2020-03-07 18:59:37 jfw: Can I assume you have an x86_64 unixlike with gcc for the install?

jfw: mp_en_viaje: I didn't think there was such a thing established, indeed, hence the question

mp_en_viaje: now then, is gbw-node a node, in the sense that it'll want to eat a metric ton of ram and piddle blockchain vomit on the disk at such a massive rate as the "too small" mb blocks require ?

jfw: right you are... I did assume the V part, huh. gbw-node requires less resources than bitcoind itself, but does require bitcoind.

jfw: Its additional disk usage is proportional to the number of addresses you tell it to watch plus the txes and outputs affecting them.

mp_en_viaje: wait, come again ? it requires less resources... ON TOP of whatever bitcoind wants ? so it's not exactly twofold ?

feedbot: http://trilema.com/2020/tell-me-how-does-it-feel/ << Trilema -- Tell me, how does it feel ?

mp_en_viaje: so it'd be fair to rewrite http://logs.ossasepia.com/log/trilema/2020-03-07#1959149 rather as "can i assume you have an x86_64 unixlike with no less than 64GB RAM, at least 1 TB HDD that must be SSD, at least two cores and, gcc. v, trb, curl etc, of which trb'd best be up to date" ?

ossabot: Logged on 2020-03-07 18:59:37 jfw: Can I assume you have an x86_64 unixlike with gcc for the install?

jfw: mp_en_viaje: you had specifically referenced the signer part, but sure, and that's why I followed up with http://logs.ossasepia.com/log/trilema/2020-03-09#1959219

ossabot: Logged on 2020-03-09 22:26:10 jfw: mp_en_viaje: do you have a specific goal in mind for Thursday's wallet work? Do you also want to use the online part (I would imagine so but could technically be done without)? If so, note that it takes about a day to scan the present blockchain once fed the address(es) of interest, and requires a TRB node. If you wish to also send the rawtx using it, as would be most proper, we'll also need that

mp_en_viaje: but doesn't it strike you as at best odd to call this an offline anything ?

mp_en_viaje: i mean... i dunno how to put it, you can do the job of producing raw tx by awk and bc in command line, and i don't mean lineS, i mean one fucking line, though it may run long.

jfw: There are two parts, one online and one offline. That awk program requires inputs. These have to come from somewhere.

mp_en_viaje: i dunno if you've noticed, but irrespective : at some point while i was travelling, there were half-hour-ish delays introduced in various processes, such as me upping myself. i even mentioned these in the logs specifically now and again.

mp_en_viaje: the ~reason~ they were there was that now and again i'd break out the red machine, which is an actual machine, and a slavegirl would sit in front of it, and type, by her pretty little hand, meaningless strings, from the black machinery connected to y'all and everything else.

mp_en_viaje: this takes time, typos occur, there's even a script to check lines individually because obviously gpg is too dumb to mark WHERE an error likely lies in a broken pad.

mp_en_viaje: now, i didn't do this because "i absolutely needed to", obviously. i chose to do it, for instance to keep my harem in good shape, and informed, connected to the world such as it is, immersed in reality as it were. i believe in such things.

mp_en_viaje: by extension from this -- and i am not mincing words when i say CULTURAL TRADITION -- i expected our adventure today would be rather similar, "this is the thing for the red machine, and these are the things you'll have to type in it : privkey, txid, txcount, so on".

mp_en_viaje: i'm not saying whether this is reasonable or not, but it is, in its context, justified as described. it's part of something, it lives somewhere.

mp_en_viaje: it provides me, for instance, the knowledge that among the things that need fixing in an "eventual pgp implementation" such as will never likely be, a typo finder'd be quite useful. i even know HOW useful, by practical, real, lived measure.

jfw: It does in that I see a typo finder would be useful; I fall short in not having one to offer now.

mp_en_viaje: that's okay, not like it's your job. but your offline wallet is ~maybe~ problematic, especially if i'm correct in not comprehending how is it supposed to actually do any useful work for a secure system.

jfw: THough the whole implementation needing to be typed manually in one line of awk/bc is perhaps the bigger block atm.

mp_en_viaje: it's one thing to say "well mister... no secure systems made before this date are practically useful anymore, because they must include this mb, and so it's practical to make NEW ones, including it". it is ANOTHER thing to say "your secure system must actually be always-on connected to a net interface and via trb at that"

jfw: ah, sure; but I'm perhaps lost at where you're seeing the node as part of the secure system, because yeah, not so secure if net-connected.

jfw: I suppose I'm the one making assumptions then. Should I have started a node up by seeking again to establish which pieces you needed?

mp_en_viaje: well i dunno. let's see, again. what i'd like to do is produce a broadcastable tx such as to pay the fellows in question out of the pile of coin i have.

mp_en_viaje: if this reduces to "you must do your computation on a machine wtih a node running" i am not interested -- it's definitely no different, and self-evidently mroe expensive than using a -- say -- web wallet.

jfw: Alright. What I have is a program that could be included on a secure machine built now, with which to do that computation, and a companion program that can suggest what you might use for its inputs given an online machine.

mp_en_viaje: anyway, thanks for going through it with me ; i hope the exercise might've been informative to you as well.

jfw: The main fusing as far as I see is that it can't even be established what your pile of coin is without trusting at least one online machine

mp_en_viaje: i thought the pile was established already ; or else what is the gasp in http://logs.ossasepia.com/log/trilema/2020-03-07#1959145 ?

ossabot: Logged on 2020-03-07 18:56:30 jfw: Moving on though: http://logs.ossasepia.com/log/trilema/2020-03-07#1959066 - I'm honored; was spinning a bit on "but what if it fucks up and zaps more coin than I've ever laid hands on??" but yeah, can't wait for another life, it's about time

mp_en_viaje: the problem of security is distinguishing public from private information. the fact that address X holds Y is not even controversial, but eminently public info

jfw: If it is acceptable to you to install new software on a machine with the private key, and to trust the inputs from somewhere online (possibly correlating from multiple sources), then we might have basis to continue. (catching up...)

mp_en_viaje: jfw, there's nothing to trust. what am i trusting ? that indeed address x holds y ? this isn't the sort of thing that requires trust

mp_en_viaje: in other words, i think your premises, not here, but in general, your ~manner of thinking~ is broken, in that you confuse very unlike things.

jfw: mp_en_viaje: the gasp reflects that there are ways to mess up even the offline signature computation.

jfw: diana_coman: not as I'd conceived it but seems that's what we're trying to determine

mp_en_viaje: but yes, the principal problem with "self-baked" tx is that it is in some cases possible for someone to re-write the tx such as to, for instance, take your inputs as fees for themselves.

mp_en_viaje: as an extension of the miner-upon-holders attack that resulted in the original bitbet death.

jfw: mp_en_viaje: yes. I don't think I'm confused in the suggested way: can't information be public but still sensitive to tampering?

mp_en_viaje: see, this is what i mean above : you take a very absolute view, which happens to be inadequate. "how is the secure machine to know" is not an interesting problem in the absolute sense you contemplate it.

jfw: adds tamper resistance, unless one is surrounded by a sybil. This is interesting though, re absolute view

mp_en_viaje: yes, "if one is surrounded by a sybil". do you know the story of the man who made a submarine so strong, it'd have withstood even falling off a cliff ?

jfw: er I mean, the absolute view point is interesting, which I guess is that ...the sybil case is not interesting?

mp_en_viaje: hitler essentially lost the war by making ever bigger tanks. "bigger is better". wel... it sure as fuck makes a better target...

mp_en_viaje: so don't you try solving the grave problem of the user being entirely surrounded by a sybil inside the secure code. let him worry about that outside the box.

mp_en_viaje: otherwise you will find yourself stuck trying to somehow make the red machine also include lasers, "just in case it needs to protect the owner from rubber hose cryptanalisis".

mp_en_viaje: this premise is contrary to the other premise, a secure system is specifically one that doesn't do everything.

jfw: It does not, in the sense that you can supply the inputs from whatever source you wish.

mp_en_viaje: alright. so then it's really gscm and gbw-signer that i want, the first being what, your hand-rolled mathlib ?

mp_en_viaje: i dunno why more people don't read more code. i swear to god usg-sponsored "comedy" is nowhere near as funny.

jfw: I confess I mostly run 'patch' by hand, but v.pl modified for keccak or any other supporting same should work.

mp_en_viaje: i'm not specifically trying to be an asshole here, even though i apparently manage splendidly regardless. do you see what might make a different me throw up his hands at this juncture ?

mp_en_viaje: invented whatever, iirc i just recognized the importance of shit other people came up with. but more's the point : you sit there "ready for sex" but you don't know where your zipper is or how it opens ? "any way!!!" ? wtf how are we to satisfyingly copulate here ?

mp_en_viaje: i'm willing to pay a visit, but... can i have like... a chair ? because no, "sit anywhere" isn't really an equivalent

mp_en_viaje: i guess it became fashionable in grenwich village at some point, just a pile of indistinct rags and "sit anywhere" -- but i confess i never sat. i always left.

jfw: well there is http://ossasepia.com/2019/11/10/v-tree-and-v-starter-v2/ but looks like there's some fix not yet incorporated into the starter, meaning... circular dependence on V to get the fix

jfw: diana_coman: I'm going by the first three comments there, starting from spyked, http://ossasepia.com/2019/11/10/v-tree-and-v-starter-v2/#comment-7012

diana_coman: I think it states quite clearly what it presses to; is that not enough for what you need?

diana_coman: jfw: why don't you take the zip, do the test run and then you know and can tell ?

mp_en_viaje: v has an intrinsic bootstrapping problem we never actually resolved, doesn't it.

diana_coman: it does at that; and my "starter" thing there is precisely a stopgap; that comes...miraculously handy at times but still apparently not worth checking upfront or something, huh.

jfw: mp_en_viaje: since it looks like I've some zipper debt to catch up on here, I'll ask - do you have GNAT available on the system you intend to press on? Because that's an indirect dependency here, but I think I can avoid it if need be.

diana_coman: jfw: listen, you do a full test-run of everything so that you can properly guide someone step by step and you know 200% what is required and at what point.

jfw: I mean, I could easily say "use this gpg command and look for this message to verify, then if that succeeds run this patch command", I tested all that, it will work, but I don't perceive it to solve the problem at hand.

jfw: re GNAT, it occurs to me the better question is not "what does mp_en_viaje have handy" but "what am I going to test and require of the user and support by explaining at whatever level necessary?" Which is not a question for mp_en_viaje at all as stated.

jfw: mp_en_viaje: would you like to schedule a continuation on the wallet effort (on which I've quite appreciated your role) once I've got my V ducks in a row? I can offer the same time Tuesday.