#Eulora log for Sunday, 2019-11-24

Latest Log | Search | Eulora Auction Price History | Eulora Wiki

« Friday, 2019-11-22
Or choose a date:
Monday, 2019-11-25 »

*** Quits: diana_coman (~diana_com@eulora/CTO/diana-coman) (Ping timeout: 240 seconds)00:03
*** Joins: diana_coman (~diana_com@eulora/CTO/diana-coman)00:05
mircea_popescuyo diana_coman11:29
diana_comanyo11:29
mircea_popescuso how are things standing with eulora dev ?11:30
diana_comanre server you mean?11:30
diana_comanI have it running locally perfectly fine11:31
mircea_popescuthis on protocuntoo ?11:32
diana_comanmircea_popescu: yes; protocuntoo on intel atm (because that was the last testing machine really, could equally well turn on the protocuntoo on amd for the same money)11:32
mircea_popescucool. and what are you hacking on ?11:33
diana_comanon server side: extracting the core (ie the already defined and implemented part of smg comms) and then porting on that bits by bits the useful stuff only (ie getting rid of all the cs in the process); seems we are quite there to do this really.11:34
mircea_popescuthat's splendid.11:34
mircea_popescuwhat's the eta ?11:34
diana_comanon client side: all this testing of different computers revealed that graphics are entirely unreliable really - something will work like this on that and different on the other and it's driving me nuts.11:35
mircea_popescuheh. article baking i imagine ?11:35
diana_comanmircea_popescu: for an initial extracting of *only very basic core* aka mainly key management (creating accounts for new registrations, sending S keys etc)11:36
diana_comanthe eta for the above would be end of this week really, I have it quite clearly mapped.11:37
mircea_popescuah joy11:37
diana_comanmircea_popescu: on client side further I dug (had no choice but to do it and I'm still hung up over it) into shaders too because a *lot* of trouble seems to stem from there11:38
diana_comanand it's probably the shaders I'll need to try and explain a bit first11:38
diana_comanfor that matter the whole variations otherwise are still - to me at least - rather..inexplicable, just nuts.11:38
* mircea_popescu listens11:39
diana_comanah, can do it here, why not11:39
diana_comanCS calls "shaders" snippets of surface-definition basically11:40
diana_comanso the idea is that you can define types of surfaces eg cloth, water, terrain11:40
diana_comanand reuse parts or all of them11:40
diana_comanwhere "defined" means a lot of different things really11:40
mircea_popescuright11:40
diana_comanincluding a number of rendering passes and what happens in each11:41
diana_comanand lighting (to some extent, not fully)11:41
diana_comanand moreover, in principle, parametrised ie having variables11:41
mircea_popescuin the sense that in this fictitious world, lighjt ~emanates from the seen object~, as in the old platonic model.11:41
diana_comanmircea_popescu: it's meant to be reflected light really11:41
mircea_popescualright11:42
diana_comanbut angles of the surface at that point (hence, direction) and otherwise properties (mainly two are used aka diffuse & specular)11:42
diana_comangetting back to the variables, those are then to be set from code11:43
diana_comanwhen you decide to use that shader on some object (of any sort)11:43
diana_comanthis much would seem relatively ok and sane but then:11:43
diana_coman1. those variables can be overwritten in fact on various levels (material comes to mind first)11:44
diana_coman2. the shaders themselves consist in a bunch of xml files that *make use of one another* in a glorious spaghetti11:44
diana_coman"for code reuse!!"11:44
diana_coman3. the internal implementation in CS using those shaders is through... plugins11:45
diana_comanand it's not even clear which ones exactly because apparently different shaders by different plugins or even possible parts of a shader require that or the other plugin11:45
diana_comansome also rely on external dependencies eg cg11:45
* diana_coman takes 1 min break to calm back down.11:45
mircea_popesculol11:46
mircea_popescuso basically you're seeing a lot of benefit from cutting out some of that expensive if useless "optionality" of the fauxly constructed ourdemocracy kind11:46
diana_comanhell yes, but the trouble is how the fuck to cut it and still have something working11:47
diana_comanespecially that if I take a knife to graphics, there won't be anything left and I'll still cut some more if it's after my heart, lol.11:47
mircea_popescuand, conversely, its cost and disutility perhaps constitutes a significant explanation of why so little movement on community dev side11:47
diana_comanah, I didn't mention that there are also at least two types of ...shader "compiler" in use11:48
diana_comanand I can't say I fully get what the difference is11:48
mircea_popescuthat they're the claims to "immortality" of two very different, subjectively, groups of fare idiots.11:48
diana_comanmircea_popescu: certainly.11:48
mircea_popescuhttp://trilema.com/2019/forum-logs-for-27-oct-2013#945020 << meanwhile in blasts from teh past. naggum lulzing at teh faretard.11:48
diana_comanbasically we have the S.MG discussion in #e for once as refugees from #t deluge, lolz.11:49
diana_comangetting back to further graphics madness since we are at it:11:49
mircea_popescuwait11:50
diana_comanall that CS dance with factories (for reusability!!!) and instances and wrappers and shits11:50
diana_comanok, waiting.11:50
mircea_popesculet me quote that whole thing, it's fucking beaUTIFUL11:50
mircea_popescu"the more insane drivel I read from proponents of Free Software and Open Source, the more convinced I am that this once noble movement will go away and not even become a footnote in the history of mankind when their real goal has been achieved: the destruction of Microsoft."11:51
diana_comanand " I'm bothered by the11:51
diana_coman  fact that stupid people don't spontaneously combust, which they should."11:51
mircea_popescu"some day, I hope you will start to listen to people who have worked with Free Software for the better part of a decade and have wondered (1) why people get much enthused but then leave disappointed and disgruntled, (2) why free software programs gain a very high quality as long as the goal is very clear, and (3) why it then goes on to accrete crappy features nobody needs but are fun to add by the less competent people w11:51
mircea_popescuho are unable to accept that an idea has fully matured.  (watch GNU ls acquire "human" sizes, for instance: multiples of 1000, instead of 1024.)11:51
mircea_popescu"11:51
diana_comanahaha, yes11:51
mircea_popescudude is so fucking on point. and this was, mind you, 90s, ie way before i for instance saw or comprehended any of this, and before by a stretch, not days not weeks. decade+11:52
diana_comanyes, he is and for all the good it did to him.11:52
mircea_popescubut this is exactly what's happening : two compilers for shaders where even one being needed at all's a stretch because of the deluge of humanity, base, useless, [http://trilema.com/2013/the-disadvantage-of-teaching-people-the-alphabet-and-nothing-more-is-that-you-have-to-somehow-put-up-with-a-bunch-of-retards-that-can-now-express-themselves-in-writing/][literate-only] whom clinton lied to into the belief they belong in11:54
mircea_popescutown11:54
mircea_popescuor whatever, ceausescu lied the same lie, it's this post-war sovietism of telling barn animals they belong in the city.11:54
diana_comanwell, the saying went to lower "the barriers to entry" so that "everyone can code", no? there, everyone DID code, isn't that great.11:55
mircea_popescuthese aren't human in any sense, but importantly ~they wish to be~ (and their bar to being is the holy "nobody can accuse of not"), and so they... make things. they "contribute". watch your city walls decay while they get [http://trilema.com/2017/the-goodbye-girl/][new drapery].11:55
mircea_popescudiana_coman, no, it's terrible.11:55
diana_comanit is.11:56
mircea_popescuanyway, the ideological excursion out of the way, let's get back to business : lotta cancercode in there you're cutting out. this is right good an' proper, let fare, poettering, althauser & co be ever remembered by the humorists and them only. nothing wrong with that.11:56
diana_comanmircea_popescu: the trouble is now wtf am I to cut out because those "shaders" are not even directly in cs code, they are a ton of interlaced xml11:57
diana_comanand then some more, what do I do, I go and cut out parts of CS?11:57
diana_comanor how exactly?11:57
mircea_popesculet me work by metaphore.11:58
diana_comanI mean: I get it, once cleaned, there will hopefully be people even wanting to do something with it; (I know I would rather do something with the clean client); still, getting to that clean...11:58
mircea_popescusuppose theres a foot between the wall and the computer, and suppose there's a quarter mile of incomprehensibly tangled power cord uniting the two.11:58
mircea_popescuthe computer works, but the tangle's the tangle.11:58
mircea_popescuso what man does is, man goes to store, buys two foot cord by spec, that fits both plugs, in wall and in machine, then turns the latter off, replaces the tangle and turns it back on11:59
mircea_popescucan you cut through the mess by producing minimally viable straight path ?11:59
mircea_popescu"minimal, viable" not "minimally viable", i mean11:59
diana_comanlol11:59
diana_comanmy current understanding is that the tangle is half-way into the machine aka CS too; hence my ????12:00
mircea_popescu(viable, by the way, literally means via-able, that can be coursed, run through, walked upon. and has, for lo these 2500 years since via flaminia)12:00
mircea_popescudiana_coman, as in there's no clearly delineated plug on the machine ? more like a placenta/umbilicum sorta thing ?12:00
diana_comanmore like 1001 ventouses where various bits attach, as far as I can see, myeah12:01
mircea_popescuso pick one of the 10001 holes competing for attention ?12:01
mircea_popescuit's how men mate after all.12:01
diana_comanhmm, not sure things are clear there somehow; let me see from the other end: do you suggest I dig in CS and figure out how to directly get from it what I want it to do, ditching all those shaders xmls and whatnots?12:02
mircea_popescuyes12:02
mircea_popescui don't see there's any other approach available if the matter's to be approached at all.12:03
diana_comanworth a try at  least; because the alternative trying to patiently disentangle that kraken of a mess will certainly take longer.12:03
mircea_popescumoreover, it's likely to produce very little in the way of upside.12:03
diana_comanthat is true12:03
mircea_popesculike disentangling a ftp servlet wittly written in php. why the fuck would i want the parts.12:03
diana_comanthe initial idea was that "minimal change"12:03
mircea_popescuthat it was, but this seems the minimal cut here.12:04
diana_comanhonestly, it probably is and moreover I would certainly much, much rather cut it out than not; but I still can't say 100% what is fully in CS and basically this is pretty much what I need to start eating now if it's to go that way, not much option otherwise12:05
mircea_popescuso basically your "as the river drives" path here is, move on to fixing client, remove cruft atop gfx engine ?12:06
diana_comanand the fact that the guy wrote it in a few months + added all sorts by all sorts afterwards is more of a downside for "full comprehension in 1 month" than anything12:06
diana_comanmircea_popescu: it sounds more like "move on to fixing client, extract core of gfx engine"12:06
mircea_popescui mean, extract minimal server first, and dump cs etc off sever that way. this to be done by end of nov ; and then, move to simplifying cs on client side, as discussed.12:07
diana_comanyes; sigh.12:07
mircea_popescuthis seems fine to my eye.12:07
mircea_popescudone... this year ?12:07
diana_comanahahaha12:07
mircea_popescuwell if not, we talk again of it come january, huh.12:08
diana_comanand then I quit12:08
mircea_popescugotta have goals in life12:08
diana_comanend of nov yes; then it's December and no, I can't promise I'll fully extract core of CS mess etc in one december and coming after all this, sorry.12:08
mircea_popescui'm not looking for a promise lol. it's just the tentative estimation. we see.12:08
diana_comano.O did I miss...goals? :))12:09
diana_comanI'll try to map it out and get back to keeping it well reported on12:09
mircea_popescupurfect.12:09
mircea_popescuaite, this is good then.12:10
diana_comanthis last month with all the pizmess and ulterior fallout on all fronts did make also a mess of reporting since well, all sorts popped up instead of what was in the plan12:10
diana_comanthanks.12:10
mircea_popescumyeah, i'm aware. it's ok, though, we're not in a distressed state here.12:11
feedbothttp://ossasepia.com/2019/11/24/some-questions-for-early-mornings/ << Ossa Sepia -- Some Questions for Early Mornings14:38
diana_comanlemme cite from CS manual for the ages: "Crystal Space is a highly modular framework. As such it is more complex then many other game engines or renderers you might encounter. The advantage of this complexity if that it offers you a lot of power in what you do and how you use it."17:16
diana_comanor such was the dream there.17:17
diana_comanto add from only 2 paras down: "the components and libraries are *more or less* independent of each other" (emphasis is mine).17:18
mircea_popescuheh21:22

« Friday, 2019-11-22
Top o' the log
Monday, 2019-11-25 »

Generated using logs2html.py 2.14.0

#Eulora log-o-tron maintained by lobbes