Hide Idle (>14 d.) Chans


← 2020-05-15 | 2020-05-17 →
diana_coman: jfw: the raw log exporter is something on my list to add and it's *no different* from the old stuff which is itself a dump of raw data, so nothing to do with mpwp or anything else
diana_coman: billymg: the "why" in the shortest and calmest manner I can muster: because it's not needed; it's an additional headache without any benefits and by headache I mean - a whole pile of *additional* stuff to maintain (that flask thing surely was mentioned in the logs but note that it requires in turn a whole set of ???, look at it sometime); the lack of permanence ultimately because look at it, if I stop logging a channel, I'm stuck ...
diana_coman: ... with either breaking down everyone's links or otherwise keeping the dead channel's logs *in the way* instead of basically still where they were but correctly a matter of the past
diana_coman: and for the record, I don't think it's the right direction there this "why not struggle with that shit some more instead of making use of what is way more useful anyway"; yes, I made the mistake of using that; no, this doesn't mean that I *have to* keep throwing good time after the bad.
diana_coman: note that I'm *not* introducing a new thing, quite on the contrary, I'm trimming away stuff.
diana_coman: and answered the comment on the blog too as apparently there were quite a few things to clarify.
diana_coman: for easy follow up, the answer
feedbot: http://bimbo.club/2020/05/work-report-5152020/ << Bimbo Club -- Work Report - 5/15/2020
cruciform: Disappointingly, the horse is usually better-looking than the rider http://logs.ossasepia.com/log/ossasepia/2020-05-09#1025859
ossabot: Logged on 2020-05-09 17:36:02 diana_coman: maybe cruciform is actually more knowledgeable re horses in the UK at least.
diana_coman: ahaha cruciform, that's the proper take on it, channeling last of the dandies muskham and all that.
cruciform: diana_coman, are you referring to this?
diana_coman: cruciform: neah; galsworthy's forsyte saga /continuation (iirc it was on top of/after the end of the main saga as such); and you know, "this" is not very helpful to link on - put the name of the thing there, way more helpful.
diana_coman: muskham was passionate about horses; by contrast, he was polite and respectful to ladies, lol.
diana_coman: he could appreciate a lady too - to the extent that she showed a good eye for the horses!
cruciform: lol - seems English women
cruciform: have been ugly for a while, then
diana_coman: well, supposedly there used to be at least something for the "English rose" but what can I say there.
cruciform: they look more like bags of fertiliser, nowadays
diana_coman: lol, at least that promises good growth!
diana_coman: but well, go where the beautiful women are, what.
cruciform: and an upcoming 2 for £10 special
cruciform: "I don't like where I live" "so move!" didn't occur to me for over a year - weird
diana_coman: you know, this thing with "3 for 2" and so on - it took me a lot of time to get my head around; why the fuck would I buy more than I wanted to buy in the first place; and why the hell does it even still end up more expensive than e.g. bigger packs at times (only yeah, had to calculate for 5 minutes to figure it exactly, what with the weird measurement system and all)
diana_coman: it does sound...weird, indeed.
billymg: diana_coman: thank you for your reply and for patiently explaining the purpose of the tool/project. i think there was a lot of context i was missing but now i understand it. left a reply on your blog too
diana_coman: billymg: answered there - it's indeed as you summarised it in your last comment; and no worries, way better to discuss and clarify matters + now they are on record so handy to give the link to anyone else who might come in wondering "why"
diana_coman: ftr and in case it's not clear, there is an argument that can be made that the old logger has a salvageable part, namely strictly the irc-frontend, if one is so inclined; it's really the web side that brings in the whole shit and trouble.
diana_coman: (well, assuming that python itself is not classed as full trouble anyway, which is debatable.)
jfw: ^ or psql
diana_coman: possibly with irc-side only, the switch to whichever one prefers can be made relatively painlessly; not that I really tried it but normally it shouldn't be a huge pain (unless it's my optimism speaking again)
jfw: my own bias (as not-a-databases-professional) would be switching mp-wp to work with psql and using that only (shouldn't be huge pain!!)
diana_coman: ahaha; why?
jfw: for instance because psql had grownup features like transactions from early days whereas mysql had them bolted on by oracle
jfw: my recollection is the traditional MyISAM storage engine is not crash-proof.
diana_coman: I really worked with people who were die-hard supporters of psql and moreover I used psql myself for years back in academia under the exact "it's more robust!! and faster!! and not-bolted on!!!!11"; having afterwards equally worked with people who found mysql the better of the 2, I can say that ...it's the people, not the tool (ie it depends on how well the operator knows psql/mysql).
jfw: I also liked that it did unix socket authentication out of the box without "plugin"; why should I need password auth when it's all on the same box
diana_coman: iirc the whole psql vs mysql has already had repeated discussions in #t too with the conclusion ~likely there will be both
jfw: also stronger type-checking comes to mind (though perhaps that was only compared to sqlite)
diana_coman: and yeah, I can see it; only atm there's no tmsr anymore and so the "both" is more like "as many and as diverse as each wants because ultimately they are anyway on their own"
jfw: would have needed a dependable lord of the database in order to benefit from picking <small number>?
jfw: nah that doesn't sound right.
jfw: in order to benefit from picking "supported database" rather. There's still benefit in reducing the proliferation, seems to me.
diana_coman: jfw: no, not at all ; and tbh I have no idea where this idea of "lord of X " came from; yeah, as I said, I did read jurov's chan log so I saw trinque's line with it but I'm still scratching my head as to where it came from
diana_coman: yes, supported database is different indeed
diana_coman: and the point is supported by ...which sovereign (most pointedly not "which lord")
jfw: weren't the minimal requirements to lordship 1) attendance in the forum and 2) the fief - something important that you owned / took care of?
jfw: and many of the titles were along the lines of "lord of X"
diana_coman: jfw: yes; that doesn't translate in any sane way to "lord of database"
diana_coman: yes, you might notice that it wasn't lord of mysql, lol
diana_coman: arguably of the whole set mine was even the most "specific" but if you look at just what eulora comprises....
diana_coman: possibly the most concrete rather than specific
diana_coman: trinque: out of curiosity, how do you even go from "lord of that realm" to "lord of pulling this lever"?
diana_coman: jfw: anyway, to get back to your earlier statement - I agree that there is still benefit in reducing the proliferation, sure.
diana_coman: for that matter there is still benefit in a small number of choices really (and mainly for newcomers since they would not end up wasting time "how do I choose")
diana_coman: but what can poor benefits do on their own - as history amply shows, not all that much, no.
jfw: ok. What sort of realms was the lordship about then? A set of duties + whatever knowledge or capital required to execute them?
jfw: that seems possibly another way of saying "this lever" though.
diana_coman: ugh, no;
diana_coman: thing is - I distinctly recall it was in #t discussed even with examples from history so I'm not fully sure what can I add precisely to make that clear since apparently it was about as clear as mud, huh.
diana_coman: the core point missing in the above though and *precisely the opposite of "pulling this lever"* is exactly... making own thing that one therefore owns, how to even put this clearer; NOT "awaiting for the duties + resources" but building one's own stuff and therefore being the lord of it, ugh
diana_coman: and myeah, I can see this "waiting patiently and all-willingly for the duties+resources" - that's how it does NOT work.
diana_coman: there's a difference between executing vs coming up with the thing to execute basically; the lordship was precisely meant to recognize the ability and deed of the latter, not the former.
diana_coman: sure, quite often and esp in the beginning one does both, but the execute part always comes after the other and there was a huge lack of that other first and foremost.
jfw: struggling for what to ask, still pretty confused I guess. So there's no answer to "what sort of thing is one lord of" because it's about each coming up with his own "thing"?
diana_coman: jfw: if you ask "of what sort of thing" , the answer would be - of a fief; if you expect that all fiefs are the same, uhm.
jfw: would an example be: at first there's no logger; someone realizes logging is important and makes one; sovereign decides it's good and grants recognition of the ability that brought it about (along with, I took it, duty to keep it working / answer questions about it) ?
diana_coman: for that matter, the very term "marquis/marquess" comes from ...lord of the march/borders, basically go ye there at the very borders and ...build them up
diana_coman: build them up and/or expand them as much as you are able to
diana_coman can't find that #t thread now, grrr
diana_coman: possibly trinque can find it since it was precisely the thread re deedbot
diana_coman: ah, finally realised why I couldn't find it, since I forgot about the mp_en_viaje handle, huh; here: http://trilema.com/2020/forum-logs-for-21-oct-2019/#2566637
diana_coman: jfw: and further note that none was ever the case of a random/assigned "pick"; it was always the only way it can even make sense - X comes in and does *on own initiative* some useful Y, the competence is recognised and as part of this recognition they get what is basically *a shot at making money and fame and whatever else they want out of it*
diana_coman: apparently though the want was wanting or did that go.
diana_coman: jfw: does that make it any clearer?
jfw: getting there.
trinque: lordship is being one of the authors of culture, and that's it.
trinque: the lordship failed because most of the folks involved were not independently wealthy and lacked the time to do said authoring.
trinque: meanwhile pretending they had by cramming w/e they could into their evenings and nights
trinque: diana_coman: re: "lord of this here lever" I mean that sitting on whatever crap one built and calling that a "fief" is not the authorship above-mentioned
trinque: also, because in the past people have assumed I'm excluding myself from statements like ^, I'm most certainly not.
trinque: at any rate, jfw I already did the postgres port if you or billymg want it
trinque: d'oh, wrong lever!
trinque: http://logs.ossasepia.com/log/ossasepia/2020-05-16#1025991 << clearly I gotta find a way to denote when I'm mocking something better!
ossabot: Logged on 2020-05-16 16:45:55 jfw: would have needed a dependable lord of the database in order to benefit from picking <small number>?
trinque: anyway, that's being a technician at best.
trinque: http://logs.ossasepia.com/log/ossasepia/2020-05-16#1025997 << this definition doesn't seem to have worked, so I'd be careful about internalizing it.
ossabot: Logged on 2020-05-16 16:48:22 jfw: weren't the minimal requirements to lordship 1) attendance in the forum and 2) the fief - something important that you owned / took care of?
trinque: http://logs.ossasepia.com/log/ossasepia/2020-05-16#1026003 << the problem here is that "realm" became so term-of-art as to become meaningless, exactly the same way socialists invert meaning.
ossabot: Logged on 2020-05-16 16:52:01 diana_coman: trinque: out of curiosity, how do you even go from "lord of that realm" to "lord of pulling this lever"?
trinque: realm does not mean place in the same way sovereign does not mean person.
trinque: seems I'm late to the party, so I'll try and catch y'all earlier tomorrow.
jfw: http://ossasepia.com/2020/05/03/ossasepia-logs-for-may-2020/#1026033 - I'd love to have it, if only for safe keeping; and why not publish anyway since you did the work. (/me glares at own pile of unsung code...) That said, I'm not likely to test it any time soon and will probably end up moving toward "mysql all the things" unless a strong reason turns up
sonofawitch: 2020-05-17 00:20:31 (#ossasepia) trinque: at any rate, jfw I already did the postgres port if you or billymg want it
jfw: some further reading satisfied me for now that the aforementioned psql bias is ill-founded, as diana_coman suggests. My own experience with non-GUI databases started with mysql (back when it indeed didn't have transactions or foreign key constraints, which left a bad taste), and my production infrastructure currently contains two mysql instances compared to zero psql.
trinque: because publishing has a certain standard, namely it working. this was just a hack I hammered out one afternoon, will see if I can find it
trinque: jfw: yeah, mysql's gotten way better over the years, and most of the rep comes from the early days.
trinque: that said psql has proved to me over about a decade of use that it's extremely reliable, and so I tend to prefer it.
trinque: one needn't be totalitarian about these things imho
jfw: trinque: I'm not in a position to be anyway, but it's a matter of "do I spend my limited hours on learning multiple large things or one large thing better"
trinque: totally sensible
jfw: trinque: meant to say - publishing doesn't have to be only "finished products"; as diana_coman's often pointed out, it can simply be about what is, or what you're up to, or what you tried, or plans... I've published drafts, dumps of notes, etc., just properly noted as such, and have yet to hear any complaints about these
trinque: yup, just working on a different item to publish, not so focused on wp atm. seems billymg has it in hand.
trinque: but I think you're right, and I could write more.
← 2020-05-15 | 2020-05-17 →