Why so many Missing Assets/Dependencies

..or perhaps Auran/N3V should publish built-in content from older Trainz versions on their own DLS if they are not to be included in a new version?

And they really should fix up the problem that Hiballer mentioned. It undermines the whole obsoleting system.
 
You may have just put your finger on the problem, Deane. The whole "obsolescence" thing is beginning to snowball into a large hairball. I realize that in order to make assets perform in newer versions of Trainz that they must conform to a stricter set of config rules. That's a given. However, what is beginning to happen now is that the very same keyword (like 'bendy', for instance) used to be a BOOLEAN. My programming background tells me that "0" is FALSE and anything else is TRUE. But now I am seeing "errors" in config files that flag a value of "2" as not being a BOOLEAN value. This is due to a change in error scanning from one version of Trainz to the next.

But, once again, this isn't addressing the overall problem of "obsolescence" and what can happen to a perfectly good asset when a new version of it appears "out there". My concern is how the devil does the DLS know there is a new version, but is unable to find it? Who, or what, tells the DLS?

Bill
 
You must have a slightly different programming background than me.

I usually do it using binary methods (0=off/not present,1=on/present).

Shane
 
Yes, I thought it was a binary thing too, but I'm not a programmer. Maybe it's one of those Brit/US cultural differences like date formats...

Anyway, there is something seriously screwed up with the DLS which must be contributing to the 'missing assets' problem.

I had always assumed that Content Manager reflected what's on the 'Black Page' DLS once it had caught up with updates. However I am now certain it doesn't. The following kuids (a Volvo wheel loader by Swedish creator Sealpower) are examples of some TS2009 assets that are on the Black Pages, but I can only get by FTP, one asset at a time.

For me (maybe only me?) Download Helper and CM just don't recognize these kuids. Why?

PHP:
<kuid:364762:1015>
<kuid:364762:1016>
<kuid:364762:1023>
<kuid:364762:1025>
<kuid:364762:1026>
<kuid:364762:1027>
<kuid:364762:1028>
<kuid:364762:1030>
<kuid:364762:1031>
<kuid:364762:1032>
<kuid:364762:1033>
<kuid:364762:1034>

I've activated a new First Class Ticket and done a database repair, and can even download more recent assets by the same author using CM, but the above kuids are still 'unknown' to my CM.
 
Last edited:
I agree, Deane, with your analysis. There is an ever-increasing gap between what is presented (and found) on the black pages versus what you can find using the various flavor of CM for each version of Trainz. In a few limited instances, I've been able to "find" what I am looking for using the black pages when CM is totally unable to find it. When I download a CDP and import it, then one of my "unlocatable" assets finally gets laid to rest.

Way back in the early days of programming when machine language was king, if you wanted a Boolean flag, you used a byte (8-bits). When you tested it (8080A language here, but the principle was the same for the 6500 Motorola series also), you used a CMP A,A or simply TST A. If it came back with the 'Zero flag" set, then there were no bits set in the byte; and you thought of it as FALSE; if it didn't set the Zero flag, then there was something in it - be it a single 00000001B or 01111111B (setting bit 7 meant negative value) then you got a Boolean TRUE. Various languages built on that mostly from FORTRAN, ADA, NELIAC, to BASIC through Visual basic/C++. If you needed a Boolean value, you used a type called BOOLEAN, but it consisted of a memory "word" (be it 8-bits, 16-bits, or 32-bits). If it had no bits set - then FALSE - any bits set, then TRUE. This can be illustrated by a Visual Basic statement "IF (Value) THEN go do something". Value can be a numerical value OR a BOOLEAN value, but it still adds up to the same thing.

Class dismissed. :wave:

Bill
 
Back
Top