Why do so many people still use Trainz 2006?

2006 uses dated scenery. I have an obsession with how something looks. It helps me to suspend belief and pretend I have a real railroad under my control. I spend more time "decorating" a route than using it for trains. It is possible, with work, to get a visually realistic route that offers the illusion of watching real trains in operation. I never drive a train since the view from a cab looks "odd" to me. If I want to experience braking, wheel-slip, horns sounds, keeping to a schedule, etc. I get in my car and drive into a city. UK..

Ts2012 can be partially tamed by blocking the SpeedTree DLLs in the User Data area/plugins. Then use clam1952 trees. The Replace function is a big help before blocking the DLLs. Then set the frame rate to 25. Now stutters and pauses are "reduced". I now am able to load up trees, bushes, well textured buildings (not the simple spray paints) plus lots of grass splines.

I found that sometimes you start a TS2012 route and the stuttering is beyond belief. However, if you restart the route the very few stutters become tolerable given what it was without the above changes. The initial stuttering is "probably" due to the program grabbing scenery, etc within a specified area and caching it if it can't find it already existing in the Data Cache. If you change routes a lot of stuff has to be cached. If you restart a route not much has to cached. Eliminate the SpeedTrees and it ain't so bad.

So, back to why stay with TS2006? For me the altered TS2012 is now working within reason. The visuals, which are important to me, are much much better. The latest assets are outstanding, if they are textured and not just spray painted with bright colors. The new clam1952 trees look great and bring back the Tundra look we all liked from the 2006 era. Lots of excellent splines of grass and other items are great in TS2012. Therefore, I would never return to "days of old".

TRS2006 is good enough for me, for testing on my antiquated PC ... If I hit the lottery, I will install TS10 ... and maybe even buy TS12 so that I can drive the NKP Nickel' Plate Road (miss speeled caboose) !

If not for the NKP DLC, that and only that, is driving me toward upgrading to TS12, nothing else but that is really worth :mop::udrool::mop: about !

RRSignal said:

In the vast majority of cases, the optimizations are none at best: rejecting a tag in config.txt serves no purpose.

Click to expand...

That would, I assume, be a wild guess on your part. Rejecting invalid tags has no purpose? Even if the tag is spelt wrong and the asset doesn't behave as expected? I think I would rather trust the guys who are actually writing the code to know what's best for the sim if you don't mind.

Paul

paulhobbs said:

That would, I assume, be a wild guess on your part. Rejecting invalid tags has no purpose? Even if the tag is spelt wrong and the asset doesn't behave as expected? I think I would rather trust the guys who are actually writing the code to know what's best for the sim if you don't mind.

Paul

Click to expand...

Judging by the reaction I got from time to time if I used the words "I'm a programmer trust me." I'm not certain programmers are the most trusted souls but in this particular case I would agree most assets with misspelt tags do run smoother with the tags corrected although I'm not certain many content creators even know what the tags mean, they just cut and paste from something similar.

Cheerio John

paulhobbs said:

That would, I assume, be a wild guess on your part. Rejecting invalid tags has no purpose? Even if the tag is spelt wrong and the asset doesn't behave as expected? I think I would rather trust the guys who are actually writing the code to know what's best for the sim if you don't mind.

Paul

Click to expand...

Um, no, it's not a wild guess, but thank you once again for not contributing anything helpful. Please tell me how breaking an asset, for, say, failing to include a thumbnail or a misspelled thumbnail helps performance? Based on your extensive coding experience, of course.

RRSignal said:

Um, no, it's not a wild guess, but thank you once again for not contributing anything helpful. Please tell me how breaking an asset, for, say, failing to include a thumbnail or a misspelled thumbnail helps performance? Based on your extensive coding experience, of course.

Click to expand...

No thumbnail means you can't see what they look like. In order to be able to use something it helps if you have a thumbnail but from recent memory it isn't a requirement for TS12 if I'm creating something just if its missing in the thumbnail tag.

Cheerio John

johnwhelan said:

No thumbnail means you can't see what they look like. In order to be able to use something it helps if you have a thumbnail but from recent memory it isn't a requirement for TS12 if I'm creating something just if its missing in the thumbnail tag.

Cheerio John

Click to expand...

Keeping in mind that the thumbnail was just an example of a tag that does not contribute to performance, I agree that in an ideal world, thumbnails should be used. I love them myself. But, they have nothing to do directly with game performance or stability. I just checked my beta build of TS12 and it does error if the image file is not included, though, you're right, a thumbnail container is not a requirement yet and the problem is, as some others have noted in the past, this breaks assets where it probably would have sufficed to issue a warning instead.

RRSignal said:

Keeping in mind that the thumbnail was just an example of a tag that does not contribute to performance, I agree that in an ideal world, thumbnails should be used. I love them myself. But, they have nothing to do directly with game performance or stability. I just checked my beta build of TS12 and it does error if the image file is not included, though, you're right, a thumbnail container is not a requirement yet and the problem is, as some others have noted in the past, this breaks assets where it probably would have sufficed to issue a warning instead.

Click to expand...

Normally if there is a mistake in the file name etc for the thumbnail tag its a simple fix as you mentioned but it can indicate a missing { so its worth flagging. I recall using tafweb's tool in TRS2004 to find and fix errors in TRS2004 so nothing much has changed except these days I see slightly fewer errors on the DLS. I certainly recall having to edit every config.txt file of more than one content creator because they couldn't spell the tag names. At least now the DLS clean up means only one person has to edit the config.txt files to sort them out.

Cheerio John

I use Trainz to build every little model railroad that strikes my fancy. Its what I use the program for. Did the Blender thing last Winter. While I have a good grasp of it, drove me nuts. I didn't enjoy it. Winter is coming again, may give it another shot. With that said, I'm an end user and personal route builder. I have stacks of magazines with routes waiting to be built and operated, my fun in this.*
For those of you who want the latest and greatest, go for it. I can't argue the content details. This pushes this down to the GPU, this tag wastes cycles....... All fun and games for some, headache for us end users. I think its cool N3V wants to clean content up. Dirty pool they dropped it on the community to figure out with no help from them. PEV deserves a cut of every game sold in my opinion. Us laymen would have been lost without him. This content clean up should have been done before new versions of the game sold. All they did was cheese us end users off to the point of tossing the game. I want to build routes, not play IT guy.
As of now I have the Mac version installed with no downloads short of Maddy's roads. Enough built in to keep me busy for years. My Windows install has TRS2006, TC with the JA files from Trainz Driver and a copy of TC with all built in removed and content stolen from 06 and Driver. Add in 5 GB of downloaded content over the years, I'm golden there as well. No need for the DLS.*
While I have JointedRail payware, it sits on a disc. No snub to them, not chasing the newer version error and DLS game to use it. Those CSX units Nikos painted up serve me well along with older RRMods stuff. I abhor Speedtrees, Profig helped, Clam put some decent billboards up but none come close to Trunda for ME.*
I dabble with the Mac version. Winter I'm into the older versions, summer is too short to play computer games. I use what works best for me. Even if I bought a new fancy computer, I prefer the older content. One of my 4 board wonders might have 5 trains running at the same time. Older versions handle it, newer does not. Newer versions stutter like crazy and the glass look of content is distracting. The old CP Geeps are fine as I sit at a signal waiting for them to pass.*
Two things. How do you turn off autosave that puts these asteriks in my posts and if anyone has the original Ohio Southern track plan from 1976...........

RRSignal said:

Um, no, it's not a wild guess, but thank you once again for not contributing anything helpful. Please tell me how breaking an asset, for, say, failing to include a thumbnail or a misspelled thumbnail helps performance? Based on your extensive coding experience, of course.

Click to expand...

Hmm, let me give one hint, processor ticks. Everytime the trainz engine looks for something that isn't there, or is misspelled, or in case of thumbnails, the tag has just a name to please previous versions without an actual file present, a performance hit is taken during runtime because of this wild goose chase.

Greetings from sunny Amsterdam,

Jan

johnwhelan said:

Judging by the reaction I got from time to time if I used the words "I'm a programmer trust me." I'm not certain programmers are the most trusted souls

Click to expand...

Hey, I resemble that. I used to be a programmer (software engineer, thank you) and I don't trust me.

johnwhelan said:

... I but in this particular case I would agree most assets with misspelt tags do run smoother with the tags corrected although I'm not certain many content creators even know what the tags mean, they just cut and paste from something similar.

...

Click to expand...

Well, I've done that because many times the WiKi does not explain well enough what some of these things do, or whether they are relevant in a given context.

But to answer the original question.

I loved TS2006 when it came out because, in my view, it was way better than the chunky TS04. And TS09 was better and so was TS10. TS12 would be better too if it were not so buggy in SP1. So I would go with TS10. I liked TS12 but my main complaint was the non functioning cab control in Driver (cab mode). That was fixed in SP1 but...

The only thing I liked about TS04 was the Ooom men.

Cheers

RRSignal said:

Um, no, it's not a wild guess, but thank you once again for not contributing anything helpful. Please tell me how breaking an asset, for, say, failing to include a thumbnail or a misspelled thumbnail helps performance? Based on your extensive coding experience, of course.

Click to expand...

I’ll admit I haven’t contributed much in the way of wild speculation to this thread but then you are doing pretty well on your own.

If you want to talk about thumbnails fine. If the thumbnail container is missing or the bitmap itself is missing or misnamed then the asset is broken. It doesn’t behave as the user expects and it’s an error, no question in my mind.

You claim that if the thumbnail can’t be found then it has no impact on the sim but I’d like to know how you know that if you don’t have access to the source code. You may be assuming that it’s not used in game but this is not true as if the 128x64 art file is missing then the thumbnails container is searched to find something else it can use instead. How does this affect the performance? I don’t know and neither do you and no amount of pretending you do know will change this.

Do you know how the thumbnail will be handled in the next version? Maybe it will be used more extensively in game, but only the programmers know that and they are best placed to define what is important and what isn’t.

As for my programming experience it is indeed extensive but again there is no way you could know that, so that is another instance of pretending to know more that you do. Actually I wonder why you are doing this, is it to somehow make yourself appear important? That’s certainly the only reason I can think of but it’s not working, at least as far as I’m concerned. All this is doing is reducing your credibility which will probably lead people to ignore what you have to say should you have a valid point to make which would be a pity.

Paul

paulhobbs said:

I’ll admit I haven’t contributed much in the way of wild speculation to this thread but then you are doing pretty well on your own.

If you want to talk about thumbnails fine. If the thumbnail container is missing or the bitmap itself is missing or misnamed then the asset is broken. It doesn’t behave as the user expects and it’s an error, no question in my mind.

You claim that if the thumbnail can’t be found then it has no impact on the sim but I’d like to know how you know that if you don’t have access to the source code. You may be assuming that it’s not used in game but this is not true as if the 128x64 art file is missing then the thumbnails container is searched to find something else it can use instead. How does this affect the performance? I don’t know and neither do you and no amount of pretending you do know will change this.

Do you know how the thumbnail will be handled in the next version? Maybe it will be used more extensively in game, but only the programmers know that and they are best placed to define what is important and what isn’t.

As for my programming experience it is indeed extensive but again there is no way you could know that, so that is another instance of pretending to know more that you do. Actually I wonder why you are doing this, is it to somehow make yourself appear important? That’s certainly the only reason I can think of but it’s not working, at least as far as I’m concerned. All this is doing is reducing your credibility which will probably lead people to ignore what you have to say should you have a valid point to make which would be a pity.

Paul

Click to expand...

I'm not a professional programmer, but I've done my fair share. Half of what I do these days is reverse-engineering other people's work looking for malware. Perhaps if you didn't assume you know it all, you'd realize that there are other people around here with tech backgrounds too.

Both from knowledge of coding and experience with Trainz, fixing assets, making assets, etc, what tags should/do impose a performance hit and what shouldn't/don't. If you had paid the slightest bit of attention to my original post, I mentioned that some tags - and errors involving them - do impose a runtime performance penalty if misspelled or incorrectly set. Many do not, and many (if not most) would not if default values were inserted, either by Trainz or preferably by CMP. Trainz already does this with a number of things e.g. trackdirections on a bridge or tunnel. No performance hit that I can see. As a result, enhanced optimization and error-checking will pay much bigger dividends for some tags and for others it will be minimal or nil.

I realize this will fall on deaf ears as you seem to think you have the answer to everything, but it may be of value to others.

Jananton said:

Hmm, let me give one hint, processor ticks. Everytime the trainz engine looks for something that isn't there, or is misspelled, or in case of thumbnails, the tag has just a name to please previous versions without an actual file present, a performance hit is taken during runtime because of this wild goose chase.

Greetings from sunny Amsterdam,

Jan

Click to expand...

The game searches for a tag whether it's there or not, though. So there is that penalty. Then there is the penalty of trying to find and load the file that may or may not be there. It might be more efficient either not do the search at all, or to relegate the search to CMP, in which case CPU cycles are far less relevant than runtime during the game, and set a default if nothing is specified. The former is a preferable strategy for tags not immediately usable by the game (e.g. thumbnails) while the latter is a better approach for important, mission-critical tags. In the case of misspelled tags, sure, that wastes CPU cycles, so CMP should strip that out and throw a warning or error and set a default, if appropriate; this, of course, does not alleviate the content creator of all responsibility to do his job correctly.

RRSignal said:

The game searches for a tag whether it's there or not, though. So there is that penalty. Then there is the penalty of trying to find and load the file that may or may not be there. It might be more efficient either not do the search at all, or to relegate the search to CMP, in which case CPU cycles are far less relevant than runtime during the game, and set a default if nothing is specified. The former is a preferable strategy for tags not immediately usable by the game (e.g. thumbnails) while the latter is a better approach for important, mission-critical tags. In the case of misspelled tags, sure, that wastes CPU cycles, so CMP should strip that out and throw a warning or error and set a default, if appropriate; this, of course, does not alleviate the content creator of all responsibility to do his job correctly.

Click to expand...

It's not so much the CPU cycles, which from my observations are minimal. TADDaemon is even quiet during game operation. It kicks in initially to start the game, but then sits idly on the side until needed. The game its self does very little with the CPU, surprisingly so. I thought it would be more CPU intensive, but it is not. What the worst culprit is turns out to be the constant disk I/O, caused by the game running. When the scenery is loaded, AI drivers are accessed or something is moved, the program has to go out to the disk and search for something, whether within the TDX file or in the multitude of hash directories, decompress the content within the hash directories, load the stuff into memory, then continue causes wasted time and interruptions.

Now keeping this in mind, this constant I/O is also impacting how the program threads are operating. This I think is a big part of the AI programming issues and is very apparent where there is a lot of content and other AI drivers. Seriously, do people only operate one tram on a city street, or have empty rail yards and mainlines? A few trains here and there, and a few trees and buildings don't cause a problem, but not all railroads operate in the middle of nowhere or at 5:30 in the morning. Most interesting rail operations take place where there are lots of buildings (scenery), track, and of course other trains. Trams usually operate on busy streets with lots of other trams, cars, people, buildings, and signals.

This content reading, analyzing, searching, and loading costs the game time, and time for us is what causes the stuttering and these I/O operations cause the awful stuttering, which I've said in another post is like being constantly poked in the ribs while trying to talk or do something else. The constant I/O drags down the system sometimes to a point of disgusting performance worse than that of a '486. Yes, turning off the Search Indexer from the N3V folders, disabling the network access (another issue too), and disabling the antivirus from scanning the Trainz data folders helps, but are not a complete fix. The problem lies somewhere before the system level and needs to be addressed.

John

Agreed. I was getting a bit of stuttering last night and I have all that disabled/not installed, and running on an SSD no less! I seem to recall a thread in the last month or two in which Windwalkr acknowledged this - in particular, the less-than-optimal database design - but I got the vibe it wasn't high on the priority list. I'll see if I can find that thread...

RRSignal said:

Agreed. I was getting a bit of stuttering last night and I have all that disabled/not installed, and running on an SSD no less! I seem to recall a thread in the last month or two in which Windwalkr acknowledged this - in particular, the less-than-optimal database design - but I got the vibe it wasn't high on the priority list. I'll see if I can find that thread...

Click to expand...

That would be an interesting read... I have one in the TS12 support forum. I had some crashing which I linked to my SoundBlaster card, but the research turned up my disk I/O studies.

John

Re Stuttering:
May be worth looking at jetlog.txt if getting over and above normal stuttering, for SpeedTree errors, such as failure to load shaders or branches or can't find whatever texture. Seems to be an intermittent thing that sometimes happens, well here anyway and then doesn't for a few days or weeks.

I haven't found the thread I was thinking of, but am finding your thread about the ability to add comments to be an interesting read. I missed the back-and-forth between various software folks and N3V. The long and short of the thread I mentioned previously, though, is that N3V realizes that the current database management system i.e. lots of small files, each in individual subdirectories of subdirectories, is a big problem. If I find it, I'll post the link.

@Malc - It seems intermittent for me too when playing off the SSD. I'll take a look at jetlog if it happens again tonight.

If the record I'm not the one claiming to know everything and you asked me specifically how a faulty thumbnail container could affect the framerates. I told you what I thought about it and said I didn't know for sure. If you don't want to hear the answers don't ask the questions.

RRSignal said:

The game searches for a tag whether it's there or not, though. So there is that penalty. Then there is the penalty of trying to find and load the file that may or may not be there. It might be more efficient either not do the search at all, or to relegate the search to CMP, in which case CPU cycles are far less relevant than runtime during the game, and set a default if nothing is specified.

Click to expand...

Bingo. You are are agreeing with me then? That's why you need error checking in CMP, so you don't do the checks at runtime.

RRSignal said:

The former is a preferable strategy for tags not immediately usable by the game (e.g. thumbnails) while the latter is a better approach for important, mission-critical tags. In the case of misspelled tags, sure, that wastes CPU cycles, so CMP should strip that out and throw a warning or error and set a default, if appropriate; this, of course, does not alleviate the content creator of all responsibility to do his job correctly.

Click to expand...

I've already given you an example of where the sim might need to access the thumbnail once it is running. Again, I don't know what the code actually does and neither do you.

I tell you what, I'll continue to assume you don't know what you are talking about and put you on my ignore list. We'll both be a lot happer then.

Paul