Crawling AI Trains.

[NeutronIC]

Before saving, press Apply, then refresh a few times, wait a couple of minutes and then press refresh again - if all still show as succeeded then it should load, I have certainly never had a problem as long as I have no "failed" showing in that list, though sometimes the "failed" can be a bit hard to spot in amongst loads of succeeded ones.

Important to note that it evaluates each train in sequence - so if you press apply it starts the process, press refresh updates to show current state, press it again a moment later and you might find one that showed succeeded a minute ago is now failed - just press refresh every couple of seconds until you are no longer seeing changes.

Matt.

[Retro]

Finished well???? when will be getting it then?

They are just keeping it in some closet for some months because they are sadists and love to see us whine for it.
Or maybe someone told them that it gets better, like cheese if you age it a bit.
Or they need to wait for a suitable Friday because game software is always released on Fridays (at least I was told so).

Like the cheese reference Andi
Regards James.

[Acorncomputer]

Perhaps there will be a cheese and whine party for the release of Mk2

[Retro]

Perhaps there will be a cheese and whine party for the release of Mk2

Nice one Geoff
Regards James.

[jamespetts]

Hi, while the work for Mk2 has finished, planning for the future has not. We do have a meeting with the Kuju coders who worked on the dispatcher this Thursday.
Once we have discussed the issues we will schedule the work.
I'm sure Adam and I will keep you posted as best we can.

regards

Derek

If one stops and thinks about the implications of this in conjunction with this post ("Who are RSDL? Where is Kuju?"), the implications are quite profound. Rail Simulator was developed by Kuju. As we now know, the coding was seriously flawed (I would say catastrophically flawed), especially in relation to the routing/pathing (or the "AI dispatcher" as it is sometimes called), and operations more generally.

Instead of the normal practice amongst reputable software developers of the original developers continuing to provide support for the product, a separate company was created to provide ongoing support, RSDL. It is clear from previous posts that the resources of RSDL (especially in terms of personnel) are extremely limited. That is confirmed by the very great length of time between releases of upgrades compared with the number and complexity of the enhancements in those upgrades.

It is now clear from this thread that the people who originally coded the "AI dispatcher", as Derek calls it, were not part of what Adam described as the "core" team that became RSDL. Quite aside from any consideration of whether or not it can possibly be right that the people who designed one of the most important aspects of the game are not considered to be part of the "core" team of developers, the more startling implication is that those who were responsible for the biggest problems have not been in a position to be doing any work to fix them.This is underlined by the fact that, over nearly a year of time since the game has been released, the two software updates released for it have had little effect on the very aspects of the game that are most broken.

Not only that, but, by the fact that there has to be a meeting with the people who coded the "AI dispatcher", it is evident that the people who were designated to provide continuing support for the product (RSDL) either did not have the expertise to deal with that sort of coding to enable them ever to fix it, and/or were not given sufficient information by the original developers as to the inner workings of the very product that it is their raison d'etre to support. That also suggests, by implication, that the original source code was inadequately commented: badly commented, or uncommented, code is a hallmark of staggering unprofessionalism in software engineering.

As the original poster in this thread noted, flaws as serious as the "crawling AI trains" "totally kill" the enjoyment of the game. The purpose of the game is not just to drive trains using the default scenarios on the default routes, but for users to be creative and create their own scenarios and routes. As is quite evident from this thread, many threads like it, and my own experiences, the multitude of critical flaws in the basics of the operations make it quite impossible to derive any enjoyment from the creative aspect of the game. It would do well for both the original developer and those who support Rail Simulator to realise that those who wish to use the game in a creative way are end-users, not developers in their own right: creating a route and/or a scenario is as much playing the game as driving a train, and it is no more acceptable to have to employ kludged workarounds when setting up signalling or timetables as it is to have to enter a console window and type a text command in order to stop the train when driving it.

Putting this all together, Rail Simulator is a game that was released with serious flaws by the original developer, so serious as to undermine beyond redemption all of the creative aspects of the game, the original developers then deliberately chose to have support for the product provided only by a small company with extremely limited resources, which, worst of all, did not include any personnel who were in a position to do anything about the things that make the game so flawed in the first place, and which in any event has so few personnel such as to make an upgrade release cycle as slow as the AI trains that are the subject of this thread.

If Kuju, or Fund4Games, or EA, or whoever else is responsible for this destitute state of affairs, had the least understanding of the market for products such as this (and had the least concern for the long-term success of their product), they would realise that (1) the market for a product of this nature is limited but decently sized; (2) it is very exacting as to the quality of products - far more so than the general game-buying public; (3) that, because of the above, the basic game could never be profitable by itself; but (4) since the target market tend to be heavy and long-term consumers of content, and is willing to pay decent amounts for such content, (witness the recent boom in model railways), the way to make money is by turning out large quantities of high quality respectably priced content, but because of (2), that model will only be viable if the basic product is of the highest quality. Indeed, although those who are responsible for such things are undoubtedly too unthinking even to consider such an option, even making the basic code (less third-party proprietary technology, such as PhysX modules, which could be purchased directly, and without anything but the example assets, which are already quasi-open-source) open-source and letting the community make it better (but not withdrawing original developer support), and focussing on turning out large quantities of high-quality content (for a price) would undoubtedly be more profitable in the long-term than releasing a fundamentally broken product and then assigning insufficient resources to support it.

Without a major structural change in the way in which Rail Simulator is supported, the whole franchise is doomed to failure, despite the large and so far largely untapped market of voracious content purchasers that could make such an enterprise signifciantly profitable if only it was properly managed.

[Retro]

Well James you have certainly put us back on Topic with your usual meticulous analysis. While not agreeing completely with your last point you have certainly given us something to think about and discuss.
Regards James.

[AndiS]

I am sure that making all or part of the code open source is on their list of options, in principal.
However, there are two catches in such an action.

A) The community is quite small, and competent programmers are sparse. As far as I know, there was some movement towards writing a train simulation by the community after the dismiss of the original MSTS 2, and it never took off because of the insufficient number of volunteers. I do not follow the OpenBVE development, but it could indicate the person power available in the scene. The fact that Zusi3 is still far from release does not add to the encouragement.

Debugging an existing software package that is written by someone else and poorly documented is really the worst job you can ever get. You need a big motivation to persevere. Also, it is often more efficient to throw parts away and write them anew. However, since the debugging task is so unattractive, the programmers (volunteers) involved will always tend to the write-anew option.

On the other hand, if the number of programmers at RSDL is small (which is obvious), then the number of volunteers might arrive at such a head count (even after scaling it to full-time equivalents. But I have my doubts.

B) Those who funded the development of KRS will think long and hard about alternatives to "giving away what they paid for". Of course, it is true that the revenues from sales of the core product are minimal by now. Also, it is clear that they can only earn on add-ons after these bugs which make the programme unpopular are removed.

But while they keep the code in their hand, they maintain the chance to benefit from the release of a version 2, Auran style. KRS without the major bugs would be a really attractive product, and since they call all the patches upgrades, why not issue KRS 2 = KRS 1 + Upgrade 4 + a new route, all integrated out of the box and uniform worldwide. Some time mid 2009, Mk.4 will be ready and MSTS 2 still far.

[jamespetts]

On the other hand, if the number of programmers at RSDL is small (which is obvious), then the number of volunteers might arrive at such a head count (even after scaling it to full-time equivalents. But I have my doubts.

If, as is very likely, the number of coders at RSDL can be counted on the fingers of one hand, then even one or two additional competent people could make a significant difference.

B) Those who funded the development of KRS will think long and hard about alternatives to "giving away what they paid for". Of course, it is true that the revenues from sales of the core product are minimal by now. Also, it is clear that they can only earn on add-ons after these bugs which make the programme unpopular are removed.

But while they keep the code in their hand, they maintain the chance to benefit from the release of a version 2, Auran style. KRS without the major bugs would be a really attractive product, and since they call all the patches upgrades, why not issue KRS 2 = KRS 1 + Upgrade 4 + a new route, all integrated out of the box and uniform worldwide. Some time mid 2009, Mk.4 will be ready and MSTS 2 still far.

At the current rate, by mid 2009, we'll be lucky to get an Upgrade 3 containing night lighting glows, more 2d map improvements, fixing the bug with the sounds of the trailing HST power car, and a few fixes for the default routes. There are insufficient resources allocated to Rail Simulator to enable the basic functionality to be fixed within any reasonable timeframe.

[phat2003uk]

James, I don't think you're in any position to judge the amount of resources that the RSDL team has seeing as you are only assuming things and not going by the facts. I think we could expect a Mk3 Upgrade by Christmas following a roughly 3 month interval. This will have to cover the more fundamental issues which you have described or I think even more people will lose faith.

[jamespetts]

James, I don't think you're in any position to judge the amount of resources that the RSDL team has seeing as you are only assuming things and not going by the facts.

It is not an assumption: RSAdam posted on this very forum shortly after release that RSDL was very small. Furthermore, one can reasonably infer (note that an inference is not an assumption) from the upgrade cycle, combined with the number and complexity of enhancements in those upgrades, that the personnel available to support the basic coding are very limited in number.

I think we could expect a Mk3 Upgrade by Christmas following a roughly 3 month interval. This will have to cover the more fundamental issues which you have described or I think even more people will lose faith.

We have not had a three monthly interval so far: note that the time between release and upgrade 1 was deceptive, since work on upgrade 1 commenced long before the release. In order to have, not only a three-monthly upgrade cycle, but one that addresses the critical issues (many of which are quite fundamental to the basic structure of the simulation itself), the output of the coders at RSDL would have to increase from its present rate by several orders of magnitude. That will not be possible without an increase in resources. I agree with the last sentence, however.

[anthonye]

I think James has put it correct and a good read, Richard say's at the end of his post "I think even more people will lose faith".

I think people already have with all of the fixes that was supposed to come, and the questions being totally ignored bu the support. Some people asked what was going to be in MK2, again no response, due to the small amount of fixes in total and fixes for default routes, I think the clever people here could have done that, me not one of them.

This forum seems to be used more by the support than there own forum, which I find odd, but there again there is more knowledge here.

Anthony

[AndyM77]

This forum seems to be used more by the support than there own forum, which I find odd, but there again there is more knowledge here.

Anthony

I'm not surprised, the forum software over there is awful!

[Acorncomputer]

Hi James (Petts)

A long and detailed analysis of the situation but seemingly a lack of understanding of the real world.

It is no secret that RSDL are a small team and have taken on the task of developing a part completed Railway Simulation program through long hours and hard work and possibly not that much reward in monetary terms at the end of the day. Small businesses do not have the resources to do all the things that some people expect and my view is that RSDL have done a splendid job so far and will keep on developing and improving this simulation with our help.

Whilst your analysis may be accurate it is just an analysis and is basically old hat and can be filed for future reference if anyone is interested.

We are not talking about an expensive professional Railway Simulation, we are talking about a game that does not claim to be perfect. Anyone expecting more than that, does not have to use it and can move on to something else.

In fact there is a tremendous support for this simulation and by giving RSDL practical feedback they can repair and improve within their resources and finance. You seem to have ignored the clear dedication and passion that RSDL have for this simulation and that is the key element missing from your analyis. Declaring that the Francise is doomed to failure is much too black and white and shows a lack of lateral thinking and a real knowledge of the tenacity of many small businesses. Making pot loads of money is not always the driving force behind an enterprise although big business these days seems to be obessed with making money and then some more whilst having very little concern for their customers. That cannot be said of RSDL.

Enough said.

[Retro]

I think we have some hope in that the MK2 Upgrade from my understanding will enable fixes to be done that where not possible before. With the enhanced Packager we may get a number of small fixes more frequently rather than having to wait for the next major Upgrade. I believe 2 have been mentioned for next week. As a Route Builder there seems to less to fix however for people who are producing Scenarios things seem to be more complicated. I can't speak for our 3D scenery and model builders due to my lack of knowledge in this area. James has raised some very valid and interesting points which seems to have sparked a useful discussion as I mentioned earlier on but I feel we must try and be as optimistic as possible at this stage in the proceedings. Time to take stock will be a year after the release date IMHO.
RSDL seem to be fully aware of what needs doing. This in part is because of their participation in these Forums which has to be acknowledged as the best support ever given to a software project, plus the support for people who use the program just for pleasure and those who are producing Routes,Scenario's, 3D models, Sounds etc. for Rail Simulator.
Regards James.

[jamespetts]

Geoff,

I am not sure that I understand - are you agreeing that the analysis is accurate or not? I do not think that it is fair to claim that I am lacking an understanding of "the real world" simply because I have not expressly mentioned that the people at RSDL appear to be dedicated and enthusiastic. Whilst it is undoubtedly true that dedicated and enthusiastic people can acheive far more than an equivalent number of people who lack dedication and/or enthusiasm, dedication and enthusiasm alone cannot, without adequate resources, fix fundamentally broken software.

I think - with respect - that you are approaching the issue from the wrong perspective. The question is not, "are the people at RSDL doing their best?", but "is the way in which the Rail Simulator project has been managed from inception until now compatible with Rail Simulator being a worthwhile product for its target market and an economic success in the foreseeable future?". As far as the consumers who just want a viable railway simulator are concerned, only the latter question is of any interest.

I also think that you underestimate the scale of the problems. While, of course, a game will be expected to have a few minor imperfections, it is also reasonable to expect of any piece of commercial software - whether designed for entertainment or anything else - that it is stable, and that its basic functionality works properly. Neither can be said for Rail Simulator. It is fundamentally broken. The fact that the people at RSDL respond to feedback does not by itself alter the fact that the company that created Rail Simulator created it broken, and then spun off the support of the software to people who, whilst quite probably enthusiastic and dedicated, lack the resources to fix it. The reality is that we have given RSDL a great deal of feedback on what needs doing, and there has been little, if any, progress so far towards doing it, undoubtedly for precisely that reason.

Retro,

the package manager will help for content updates, but most of the critical issues are fundamental problems with the code, not problems with the assets. It is noteworthy of itself that such a high proportion of fixes so far have been fixes to assets rather than basic code, which gives some idea of the balance of expertise in RSDL.