ECS vectors v1.1.2 (by George) 19/06/2011

[wallyweb]

I'm looking for beta-testers for ECS vectors beta 5.

If they are all done, I can test in TTDPatch (latest nightlies).

[George]

I'm looking for beta-testers for ECS vectors beta 5.

If they are all done, I can test in TTDPatch (latest nightlies).

Unfortunately, since beta 5 ECS vectors would not support TTDP. There are some unsupported features and there are no plans to support them
1) long date format vars
2) more values for industry var 93
3) 32767 sprites per GRF limit (Yes, ECS Machinery vector has finally achieved this limit, and has some plans to grow even more)

[wallyweb]

Unfortunately, since beta 5 ECS vectors would not support TTDP. There are some unsupported features and there are no plans to support them ...

Too bad ... I will stay with beta 4 then as well as Pikkabird's industries and hopefully some nice new Canadian industries when DanMack has then done.

If the intention is to only support OTTD, perhaps it is time to move this thread over there so that the players will not be confused.

[George]

Unfortunately, since beta 5 ECS vectors would not support TTDP. There are some unsupported features and there are no plans to support them ...

Too bad ... I will stay with beta 4 then as well as Pikkabird's industries and hopefully some nice new Canadian industries when DanMack has then done.
If the intention is to only support OTTD, perhaps it is time to move this thread over there so that the players will not be confused.

May be... But how?

http://george.zernebok.net/phpBB/viewtopic.php?t=80

I'll do some attempts to add the code to support TTDP, but I do not expect to finish it soon

[Chico008]

me and some friend run a server with beta4 of your ECS.
we can test your Beta5 if you want to.
we run an r14004 server with Tiberius Patch Pack.

[wallyweb]

http://george.zernebok.net/phpBB/viewtopic.php?t=80

OOOH! I've bookmarked this for future reference.
Maybe add this link to the first post of this thread.

May be... But how? ...
I'll do some attempts to add the code to support TTDP, but I do not expect to finish it soon

Well, if you plan to do this then I vote for this thread to remain here. Just keep me in mind when you plan to start.

[Gremnon]

I can test them in the OTTD nightlys, with the daylength patch. If that matters.

[George]

me and some friend run a server with beta4 of your ECS.
we can test your Beta5 if you want to.
we run an r14004 server with Tiberius Patch Pack.

I can test them in the OTTD nightlys, with the daylength patch. If that matters.

Any testing is welcome. Just report me if something goes different way than wiki says ar written on the forums. Oven if you've expected.

[Chico008]

Does the parameters for the ECS are the same as Beta4 ?

where can we dl the beta5 ?

thx

[PhilSophus]

If the intention is to only support OTTD, perhaps it is time to move this thread over there so that the players will not be confused.

May be... But how?

By using the report function on one of your posts (the little button with the exclamation mark) and asking if the whole thread could be moved.

I can test them in the OTTD nightlys, with the daylength patch. If that matters.

Any testing is welcome. Just report me if something goes different way than wiki says ar written on the forums. Oven if you've expected.

However, you should be aware that with months being longer with the daylength patch, industry production is updated less frequently. I had the impression that this causes ECS vectors to overreact. If you didn't provide more and more trains for e.g. a mine the production would tend to collapse instead of stabilizing around a certain level. Playing a non-daylength game now, which does not behave as that, I noticed that this must be an issue with daylength (maybe the patch should call the monthly loops n times per month for a daylength level of n) not a problem of ECS.

George, are you planning to pull the construction vector(s) to beta at some time? I quite like the look of some of the industries in there, like cement works, lime quarry or brick works but lacking the nice features of the beta vectors like stockpiling they are not that much fun to play.

Anyway, you are doing a great job here, I enjoy playing with your ECS vectors very much and don't have the feeling that I'll soon hit a limit trying out new things.

[Gremnon]

I'm aware of that, that's why I offered to test it with the daylength patch. That, and it's an interesting challenge, unlike my recent NoAI games where I got beaten slightly...

[DaleStan]

There are some unsupported features and there are no plans to support them
3) 32767 sprites per GRF limit

You know, I'd take this a lot better if, among other things:
1) The simple expedient of "remove the nops" actually left more than 32767 sprites,
2) You did not use action A to change sprite 2 *fourteen* separate times, and
3) The comment-stripped NFO did not compress nearly three times better than any other NFO I checked. (Procedure calls FTW.)

As for vars 23 and 24, you could simulate them with only six extra sprites total. Put those six place of the 12 extraneous sprites that change TTD sprite 2.

This leaves just indu var 93, and if you're willing to be a guinea pig, I can implement CB29/35 returns 0D and 0E. I just don't have the patience to test them.

[michael blunck]

Yeah, I´ve never seen before a .grf where the .pcx file decodes to a size of 2 MB but the .nfo uses 8 MB. I know the way George likes to handle things since a couple of years, but that´s just over the top.

regards
Michael

[DaleStan]

Put those six place of the 12 extraneous sprites

By which I meant, of course, "Put those six in place of the twenty-six extraneous sprites ..."
(Remembering that changing one TTD sprite requires two NFO sprites FTW too.)

@MB: May I ask how many sprites are currently in DBSetXL 0.9? Total -- the value in sprite 0? And/or what you expect it to be when 0.9 is released?

[michael blunck]

@MB: May I ask how many sprites are currently in DBSetXL 0.9? Total -- the value in sprite 0? And/or what you expect it to be when 0.9 is released?

ATM, DBXL 0.9 contains 9894 total sprites and 5789 "relevant" sprites (shown in status window).

This is due to tight coding (procedure calls, recolouring, ..), because 0.9 contains approx. 60% more "vehicles" than 0.82 plus animations. (V0.82 contained 6645/4608 sprites).

regards
Michael

[George]

Does the parameters for the ECS are the same as Beta4 ?

No. Beta 5 has more values for parameters. See the wiki

where can we dl the beta5 ?

beta testing goes here http://george.zernebok.net/phpBB/viewtopic.php?t=80

I can test them in the OTTD nightlys, with the daylength patch. If that matters.

Any testing is welcome. Just report me if something goes different way than wiki says or written on the forums. Oven if you've expected.

However, you should be aware that with months being longer with the daylength patch, industry production is updated less frequently

Does production callback happen more than 9 times a month?

I had the impression that this causes ECS vectors to overreact.

It depends on what does the patch change. If the tick time is the same, than it may cause ECS vectors to go mad.

If you didn't provide more and more trains for e.g. a mine the production would tend to collapse instead of stabilizing around a certain level.

do beta 5 fix the problem?

George, are you planning to pull the construction vector(s) to beta at some time?

Yes.

[George]

There are some unsupported features and there are no plans to support them
3) 32767 sprites per GRF limit

Do not worry Dale, this is the easiest problem to fix. As soon as someone would fix long date problem (the most annoying part to fix) I'll fix this ASAP.

As for vars 23 and 24, you could simulate them with only six extra sprites total. Put those six place of the 12 extraneous sprites that change TTD sprite 2.

This is the most annoying part to fix. Could you provide long date support in TTDPatch?

This leaves just indu var 93, and if you're willing to be a guinea pig, I can implement CB29/35 returns 0D and 0E. I just don't have the patience to test them.

Fine! Just whistle

@MB: May I ask how many sprites are currently in DBSetXL 0.9? Total -- the value in sprite 0? And/or what you expect it to be when 0.9 is released?

ATM, DBXL 0.9 contains 9894 total sprites and 5789 "relevant" sprites (shown in status window).
This is due to tight coding (procedure calls, recolouring, ..), because 0.9 contains approx. 60% more "vehicles" than 0.82 plus animations. (V0.82 contained 6645/4608 sprites).

Just to remind. LV4 has 10242 total sprites

[PhilSophus]

However, you should be aware that with months being longer with the daylength patch, industry production is updated less frequently

Does production callback happen more than 9 times a month?

As production per month also scaled by the daylength factor, I guess so. But what I meant is, the production level is only updated once a month, no matter how many ticks that is. Right?

I had the impression that this causes ECS vectors to overreact.

It depends on what does the patch change. If the tick time is the same, than it may cause ECS vectors to go mad.

The patch changes how many ticks there are per game day (e.g. daylength factor 2 means 148 ticks per day instead of 74 and thus, a month also has twice the number of ticks). As the station rating is updated each 184 ticks as far as I understand from the source code, that essentially means that the industry sees a drastic change in rating during a month.

If you didn't provide more and more trains for e.g. a mine the production would tend to collapse instead of stabilizing around a certain level.

do beta 5 fix the problem?

Sorry, at the moment I only play with my own timetable patch and I don't want to combine it with the daylength patch as this would make testing harder.

George, are you planning to pull the construction vector(s) to beta at some time?

Yes.

I guess, if you wanted to write more on that, you would have done so. So, I don't ask any further But I'm glad to hear that it will happen at some time.

[George]

However, you should be aware that with months being longer with the daylength patch, industry production is updated less frequently

Does production callback happen more than 9 times a month?

As production per month also scaled by the daylength factor, I guess so. But what I meant is, the production level is only updated once a month, no matter how many ticks that is. Right?

Yes. But industry takes amount of cargo produced into account, that's why current level calculation should be confused.

I had the impression that this causes ECS vectors to overreact.

It depends on what does the patch change. If the tick time is the same, than it may cause ECS vectors to go mad.

The patch changes how many ticks there are per game day (e.g. daylength factor 2 means 148 ticks per day instead of 74 and thus, a month also has twice the number of ticks). As the station rating is updated each 184 ticks as far as I understand from the source code, that essentially means that the industry sees a drastic change in rating during a month.

Does this mean that production callback would happen every 256 ticks, not every 512 ticks? If so, most industries should be confused

I guess, if you wanted to write more on that, you would have done so. So, I don't ask any further But I'm glad to hear that it will happen at some time.

As you've expected, the answer for "when?" question is "I do not know".

[PhilSophus]

Does production callback happen more than 9 times a month?

As production per month also scaled by the daylength factor, I guess so. But what I meant is, the production level is only updated once a month, no matter how many ticks that is. Right?

Yes. But industry takes amount of cargo produced into account, that's why current level calculation should be confused.

I see. And it takes the production per month into account and not the production per production event? In the latter case it should see no difference.

The patch changes how many ticks there are per game day (e.g. daylength factor 2 means 148 ticks per day instead of 74 and thus, a month also has twice the number of ticks). As the station rating is updated each 184 ticks as far as I understand from the source code, that essentially means that the industry sees a drastic change in rating during a month.

Does this mean that production callback would happen every 256 ticks, not every 512 ticks? If so, most industries should be confused

Mind you that though I have looked at the code I'm not an expert on this patch. To my understanding, everything that should happen every X ticks still happens every X ticks (and ticks are as long as before). Everything that is tied to the calendar scale, like daily, all 20 days, every month happens less often (by the daylength factor, so daily runs every n*74 ticks instead of every 74 ticks).

Actually, ECS is behaving astonishingly good with daylength. To be honest, when I first tried ECS with the daylength patch I expected much worse effects. It's actually behaving so well that not until recently I realized, that ECS did react differently with daylength patch.

Anyway, I think it's rather the duty of the patch to make sure it is not breaking stuff, not that of your NewGRFs that they work with experimental patches. I just wanted to make you aware that if people try it with the daylength patch (as someone offered) problems may be reported that are rather the problem of the daylength patch.