DDL wrote:Salk wrote:Since this topic is addressed to DDL, I take the opportunity to ask him a few questions, hoping to not divert too much of his precious time from working on HDTP!
Seriously, I don't know anything about Deus Ex modding but I am a rather curious person and I am sure your answers would be of interest to many other lurkers:
1) What does your HDTP work consist of?
2) Is it you the only one working on HDTP at the moment? If not, who else, and what are their tasks?
3) Once HDTP is released, do you think it will be final or there will be updates to fix possible bugs/errors?
4) Could you estimate how many hours you have spent on this project?
5) Considering your answer to question number 4, do you ever feel regret for even having been involved in HDTP?
6) How has a biochemist ended up working on HDTP?
Dave's answered a lot of these already, and incidentally has also massively underplayed his own involvement (involvement, incidentally, which has been awesome). Christ, some of the stuff he just "magics up" overnight is...fucking incredible. Mind you, he's apparently incapable of making any texture smaller than 1024x1024, so if you ask him to rustle up a texture for a 'plain, red, button', you're gonna end up with one fucking INCREDIBLE button, but your texture caching won't thank you.
So yeah, Dave's been doing lots of textures and modelling and UVmapping (not, of course, normally in that order).
Akerfeldt, as noted, has been doing a
fuckton of stuff, and he's very much at the magical coding end of the spectrum: where I am reasonably proficient at making the shitty unreal engine play ball, he seems able to write his own software that can then do things to super-high-def-uber-models which will A)animate them with vastly less fuss than one normally encounters, and B)make them STILL appear to be super-high-def-uber-models when in-game, rather than "super-high-def-uber-models regrettably mangled by the unreal engine".
Metche, who has (understandably) been fairly distracted recently what with various job moves, and (I believe) getting married, and getting basically worn down by massive accumulated HDTP fatigue, has done tons of organising, some textures, more organising, and has essentially acted as the central focal coordinator of HDTP efforts for god knows how long (and also a source of collective motivational guilt: "you can run away and just try to pretend HDTP isn't happening and you don't have to do anything, but Metche will fiiiiind you....and ask you nicely to do some stuff").
Myself, ultimately....well: I'll answer 1 and 6 together: modding the unreal engine and biochemistry research are actually very similar, in that both are essentially (on a day to day basis) composed of mindless, mindless, fucking mindless, repetitive tasks. A child of like...10 could do my job (both in HDTP and biochemistry...though admittedly on the biochem side I don't currently have a job..but you get my point). The catch is, they could do so only if everything actually worked according to plan.
Things NEVER go according to plan. Much like biochem, where you hire a postdoc
not to do northern blots every fucking day, but to be able to do northern blots (or suitable alternative)
even in the face of major obstacles, and to know why something when wrong when it did, and what to do to fix it.
For HDTP this could be things like rearranging texture order to fix muzzleflash wiping, or if that's not possible, recoding the weapon to use different texnums for muzzleflashes. Or making sure a given texture has the right flags on a mesh. Or reUVmapping it slightly if it turns out it has conflicting texflag requirements on the same multiskin. Or readjusting the UVmap to avoid texture shearing. Or even adding a muzzleflash if one isn't on the model, and THEN doing all the above. Or in fact making a muzzleflash myself, UVmapping it, adding it, and then doing all the above.
And that's just for muzzleflashes.
There's also the fact that HDTP has taken many contributions from many different people, many of whom did not perhaps totally understand the brief, so I've also spent a lot of time taking essentially unusable stuff and working out how we can use it.
My work on HDTP is basically being willing, able, and inventive enough to get everything imported in the first place (which is actually fairly easy, but tedious as FUCK: I've kinda got that role because most people can't be fucked learning how to do it, and the others can't bring themselves to confront the tedium), then get the code references written, and account for all the clever stuff like weaponmods magically appearing on weapons when you add them, and
then to fix the myriad number of tiny irritating bugs that can crop up, which really means I need to be capable of doing a lot of things (textures, modelling, UVmapping, fucking tweaking import lines...*shudders*) to a...reasonable standard. Jack of all trades, master of none.
Incidentally, this is also pretty much what my academic CV looks like, which is not helping me much: you wouldn't THINK it'd be possible to have 'too broad' a focus in a subject as inherently niche as biochemistry, but apparently it is.
2) has been answered by both me (above) and Dave (higher above), so I'll move onto...
3) probably. I can't guarantee it'll be prompt, but I don't think I'll be leaving this forum anytime soon, so I should be able to read through lists of bugs and tackle those I can. Akerfeldt and Dave will also hopefully be around to do those I can't. Admittedly some of them might be "slightly gippy animation in model X during reload while under obscure situation Y", which I reserve the right to respond to with "ah well. Still looks better, eh?", because animation fixing is (even with Akerfeldt's magic program) still quite an arse.
4) I really really shudder to think. What gets me is how much of that time is tedium. Not OOH LOVELY LOVELY MODELS, but
"fucksake: it's offcentre by about 3 uunits"
10*delete package*
20*adjust import code offset by arbitrary number (because import code offsets and in-game offsets do not correspond in any convenient fashion)*
30*recompile*
40*reload UEd*
50*check model*
"ok, so that's gone about 5 uunits over, so if x=1500 was 8, then 3 would be..X=562"
60*GOTO 10*
...but then, fuck: I spent a year and a half doing NOTHING BUT radiolabelled RNA stuff, so I think I'm reasonably good at handling tedium.
5) Hah! I'll tell you in about a month's time.
I'm still proud of our first release, so hopefully that'll carry over.
And for the record, we're definitely on the home stretch now.
but all did look excellent! Haven't heard from anyone in ages so couldn't really say.