SimCity: Welcome to the Glassbox Simulation Engine

[packersfan 666]

I just read the story at Game#&*#...I forgot where, ha, but wow, they make it sound like the game will ACTUALLY reflect what is going on in your city, like you ACTUALLY will see what is ACTUALLY happening. As opposed to SC4 where graphs say ONE thing, but the game shows something radically different. When SC5 was initially announced I was a bittersweet feeling, knowing that we will all move on and we will have to restart the customization process (though if the tools are provided for this game, I have no doubt the SimCity community will do incredible things with SC5), but now that I read about this game engine, I am excited and really can't wait for a game that sounds like it will be more playable, more realistic, more responsive. This is of course what we EXPECT from next chapter that has been 10 years coming.

I am intrigued by the mentions of SimTower and other old school Sim games using this engine in future revivals, that would be a money maker!

Here is what I am always concerned about - realism. Yes, we want a fun, playable game, that isn't SO complicated it takes multiple masters degrees to play it, but we also want realism. We want the freeways to look and act like freeways, we want transportation systems to be completely (or nearly) customizable. We want realistic looking buildings and the ability to dictate the kind of buildings (i.e. type (through zoning), style, time period, wealth level) to be built in certain areas. This is sort of turning into a wishlist, haha.

Anyways, I am excited to see how GlassBox revolutionizes this game and makes it more than just a sandbox game for me!

[Techleo 18]

Realism is definitely something a lot of us are concerned by. As it stands, I think the Unit, Agent, Resource and Rule system really offers us a far more organic and realistic system. Instead of the pc running calculations and rending traffic as it supposes it would look, we have cars using roads.

The cars interact with there environment and traffic forms as a consequence of interactions. Rather then a pure calculation rendering.

That realism seems to pervade every little element. Every tree, bunch of grass, person, aquifer and so on all links to each other and the behaviours form organically.

That excites me beyond imagination. It also seems to me it will make modded structures have a far wider impact. Thats good in a way.

I read a bit about how the farming works and that is one thing I am thrilled with. Farms take in nutrients and water. As long as those substances remain, the farms grow. Eventually you can have a scenario similar to the dust belt. If water tables disappear and nutrients is to low, the area dies and it cascades outwords.

This makes it necessary to manage nutrients and water tables. Eventually you will have to rotate farms off land and onto new land to ensure regenration of the water table and the nutrients.

This is the sort of simulated reality that we will just see in a simple manner that looks and makes sense. Yet its immensely robust. Good stuff

[A Nonny Moose 8,262]

It certainly cast some serious light on all this. Now, if management at EA doesn't jam the product on the market to meet some stupid deadline before it is debugged, things will be interesting. Remember that Glassbox is only the simulation specification interface to the core program. Nothing has been said about the graphics side.

One thing is clear from the reddit is that the old concept of a megalopolis on one tile is out. The trick now will be to spread out. If you want a central business core, so be it, but it will likely be a separate tile from the rest of the "city". This can be inconvenient unless the load time for switching tiles is quite fast.

And since it is on-line, how much co-routine will be resident in the PC and how much on the server is not at all clear. I think it may work like a super-browser with only one destination, and much of the substance will be on the server. Only the graphics buffers and a small I/O kernel may be at the user's end. A fast line will be absolutely necessary.

The server is likely to be a mainframe or a hypercube which can only be supported by a big outfit with lots of users. I hope they have enough capital to support it during the start-up period.

Oh, and as for bug fixes, if any, since the program will be on a central server, users need not even be aware of it much.

[croxis 150]

It certainly cast some serious light on all this. Now, if management at EA doesn't jam the product on the market to meet some stupid deadline before it is debugged, things will be interesting. Remember that Glassbox is only the simulation specification interface to the core program. Nothing has been said about the graphics side.

One thing is clear from the reddit is that the old concept of a megalopolis on one tile is out. The trick now will be to spread out. If you want a central business core, so be it, but it will likely be a separate tile from the rest of the "city". This can be inconvenient unless the load time for switching tiles is quite fast.

And since it is on-line, how much co-routine will be resident in the PC and how much on the server is not at all clear. I think it may work like a super-browser with only one destination, and much of the substance will be on the server. Only the graphics buffers and a small I/O kernel may be at the user's end. A fast line will be absolutely necessary.

The server is likely to be a mainframe or a hypercube which can only be supported by a big outfit with lots of users. I hope they have enough capital to support it during the start-up period.

Oh, and as for bug fixes, if any, since the program will be on a central server, users need not even be aware of it much.

I highly doubt this. This isn't a single server persistent world mmo but small 16 player regions that communicate asynchronously. The key evidence is in the answers that the gamer can play offline from the reddit discussion and the entire slide 48 of the GDC talk. The only thing we know is that the games are saved on EA's servers.

  Edited March 9, 2012 by croxis  

[A Nonny Moose 8,262]

It certainly cast some serious light on all this. Now, if management at EA doesn't jam the product on the market to meet some stupid deadline before it is debugged, things will be interesting. Remember that Glassbox is only the simulation specification interface to the core program. Nothing has been said about the graphics side.

One thing is clear from the reddit is that the old concept of a megalopolis on one tile is out. The trick now will be to spread out. If you want a central business core, so be it, but it will likely be a separate tile from the rest of the "city". This can be inconvenient unless the load time for switching tiles is quite fast.

And since it is on-line, how much co-routine will be resident in the PC and how much on the server is not at all clear. I think it may work like a super-browser with only one destination, and much of the substance will be on the server. Only the graphics buffers and a small I/O kernel may be at the user's end. A fast line will be absolutely necessary.

The server is likely to be a mainframe or a hypercube which can only be supported by a big outfit with lots of users. I hope they have enough capital to support it during the start-up period.

Oh, and as for bug fixes, if any, since the program will be on a central server, users need not even be aware of it much.

I highly doubt this. This isn't a single server persistent world mmo but small 16 player regions that communicate asynchronously. The key evidence is in the answers that the gamer can play offline from the reddit discussion and the entire slide 48 of the GDC talk. The only thing we know is that the games are saved on EA's servers.

Well, the question is how many (up to) 16 player sessions do you think they can support? If you choose to view a member of your group's efforts while he is on-line, life gets interesting. Clearly multiple writers on the same city file opens a can of worms that is unlikely to be made available.

Start thinking in terms of groups spanning the globe, and of a session for each group. If the action on this site is any example, it could be upwards of 500 users in possibly 100 sessions. Would work fine if there were only one, or a few copies of the program text, but makes for a huge problem if each user has his own copy of the game text at the server. I think the distributed program paradigm will be the best approach. Updates could be distributed at login time.

[Ocram's Razr 964]

Let us see...

1. All cities are stored on the cloud - large servers needed

2. Resource prices are affected by all save games - all cities somehow linked

3. Comparison to Cities XL Planet Offer (500 cities per planet?, different planet for each country plus demo and test)

4. Comparison to Anno 2070 (insufficient information...please research this)

[Auriga_Ando 0]

I really like the Idea behind that glasbox engine. Especially we will now have a real rush hour in SimCity. In Simcity4 or CitiesXL you turned on the graphs and see a red road between your residential and industrial areas. That indicates you have a high traffic load there, but that high load is 24/7 all around the clock and the traffic jam in the morning just an animation.

In SimCity5, after the things I have read so fare, we will see something like this: In the night roads are empty and in the morning people start driving to work. Now you will get a high traffic load on the road between your residential and industrial area and a real traffic jam not just a animation. After a wile people will be arrived and everything is fine on the streets again, now factories starting to manufacture stuff and trucks will transport goods between your factories in your city. That will cause some traffic load again and in the evening we will have rush hour again.

This level of realism is awesome and in my opinion this will give us some interesting options to handle traffic in our cities. For example avoiding crossroads when constructing a central traffic axis in a city will have a real positive impact on how fast people can get from one side to the other side of the City. Sunken roads, elevated roads all that will have more sense with that kind of simulation.

  Edited March 10, 2012 by Auriga_Ando  

[croxis 150]

We know glassbox is asynchronous and the player does not send the save file, but just the data of what to add/change so you only need a single writer and queued message system, no different than a minecraft server where you have a bunch of people changing the same world data.

We know that the engine supports cities interacting with each other, but they have not said (that I know of) that player's can visit other people's cities.

We know from this slide that the multiplayer servers are mostly saved game holders and data relaters (hey city b! city a just sent you 20 pollution!) but it doesn't look like, at this time, it does much processing on its own. This is opposite of most online games where the client is "dumb" and the server does all the processing.

[Ocram's Razr 964]

We also know that multi-player cities are "saved on the cloud" (hosted on the servers). Depending on the hours of operation for shops and services, you might have a bit of off-peak "leisure" driving, possibly even late into the night (night clubs). Did you notice that the game is Pegi 12 when Anno 2070 is Pegi 7 (and it has war)? That means that crime is going to be simulated (is Pegi 12 rated T or E-10?). I certainly hope that all the crimes in SimCity 4 will make it to SimCity 2013, with all the features of Glass Box and the rendering engine (remember that SC4 had bank robbery, drug possession, joyrides, flashing, murder, assault, and arson among others, though some of those might be relegated to Dr. Vu, which I hope is not the case).

[A Nonny Moose 8,262]

Let us see...

1. All cities are stored on the cloud - large servers needed

2. Resource prices are affected by all save games - all cities somehow linked

3. Comparison to Cities XL Planet Offer (500 cities per planet?, different planet for each country plus demo and test)

4. Comparison to Anno 2070 (insufficient information...please research this)

A little too general. Think of it as a set of nested boxes and you should get it.

The outer container is the group of players for any one region and/or the region.

Only cities in the region are linked for effects.

There can be many unrelated groups.

A group consists of 1 to 16 players.

It makes sense to have some kind of fast data base (a B* tree) to store the data in some kind of compressed format. If there is also a fast repository of lots/models/textures, then the region need only exist in virtual memory when it is being played.

When you sign on, as soon as your group is identified the composition of the region can start. Typing is slow, so it may finish almost as soon as you are ready, because there is also image transmission time if your station doesn't have a current image. With good programming, only updates.

Remember, not all the members of your group need be on line simultaneously, so update journals have to be kept for each player in a group. This can get quite large until there is a group commitment point when all journals can restart.

There is nothing magic or new about all this. On-line systems of this type have been around since the late 1950s. How do you think on-line banking works. The ERMA system was initiated by GE in the time frame when I was working for them. When we installed new banking systems in some Canadian banks in the 1960s, the any branch banking stuff was already there, but the banks were afraid to turn it on. People still wanted to carry passbooks. In fact, some L.O.L.s still do, and have to go into a branch to have them updated.

Now, the complexity in all this comes from having several groups/regions in play simultaneously. If your region isn't active, it exists only as entries in the data base and perhaps an image stored on your local PC. However, for every region being played, you have to handle up to sixteen terminations, updating each one as needed and accepting transactions from them that cause the updates. With only 16-way considerations in any discourse, this is rather trivial. Systems these days handle thousands of simultaneous nodes. It is only a matter of being sure you have enough machine to prevent people from becoming too psychotic waiting for their results. Studies have shown in the good old days that good response time in a system like this is three seconds or less. After three seconds, some kind of psychosis sets in that results in double sends and other silliness like hitting the return button five times. The fun begins for the server when an area of the region is being affected by more than one player. Now you have to be careful about commitments and deadly embraces.

Asynchrony may make these occasions rare, but you still have to handle it, because if you don't it will surely happen. Historians of the U.S. Army Academy at West Point will know about this, and why Edgar Allen Poe was expelled from the Point. Being right doesn't always get you points if the bone of contention is the plumbing system in the barracks.

As far as server hardware is concerned, I think a couple of 1980's era AIX boxes would do the job as long as it had enough disk. These Power PC 128-bit processors are lightning fast and not too much money. The International Brotherhood of Magicians would be delighted to help. One would do, but you have that 24/7 requirement, so shadowing is needed with a guaranteed UPS.

[Ocram's Razr 964]

OK, I researched multiplayer in Anno 2070 (partly from playing, partly from Internet) and all games are saved on the players' computers. Multiplayer is simple because each map has a seed that has the same results for all players and all units are the same for all players. Therefore, Anno 2070 copies the playing field to your computer when you join a multiplayer game and only sends data on what has changed (building placement and destruction, resource balances, ship/unit fleet construction, location, and combat, etc.

I know that IBM was the leading computational machine producing (card readers, vacuum tube mainframes, beige box home computers) company of the 20th century but I thought they got out of the business with that Lenovo deal.

[A Nonny Moose 8,262]

OK, I researched multiplayer in Anno 2070 (partly from playing, partly from Internet) and all games are saved on the players' computers. Multiplayer is simple because each map has a seed that has the same results for all players and all units are the same for all players. Therefore, Anno 2070 copies the playing field to your computer when you join a multiplayer game and only sends data on what has changed (building placement and destruction, resource balances, ship/unit fleet construction, location, and combat, etc.

I know that IBM was the leading computational machine producing (card readers, vacuum tube mainframes, beige box home computers) company of the 20th century but I thought they got out of the business with that Lenovo deal.

IBM no longer markets PC hardware. However, they still sell everything else. lenovo

Big Blue decided that they didn't want to be in the business they started as it became irrelevant to what they are doing, and they consider it rather cut and dried. Besides monsters like Watson, they are also working on technologies like very high density rotating storage, and other future stuff. Their mainframe business has always been profitable, and they see no reason to abandon it.

A lot of people these days think mainframes have passed, but it has simply regrouped and gone into the big business world. You can't do everything with a PC or a network of PCs. If you need sheer power, give me a hot mainframe every time, even if it has to be cooled with liquid helium. Of course these things sell for megabucks, as they always have.

[Ocram's Razr 964]

Yes, mainframes will most likely always be needed, along with servers. Despite the fact that your average gaming PC today has more power (when using CUDA) than a mainframe of the 1990's, major companies still need servers and data centers as well as the now uniquous office computer.

[A Nonny Moose 8,262]

^ Ever seen a PC with a vector processor for doing little things like inverting matrices 1 MB in size, or handling 64 digit floating point arithmetic in hardware? There are tasks that are not suitable to a PC's tiny selection of binary operations. Imagine a machine with an 18 bit order code and three operand addressing. Had one before 1990.

  Edited March 12, 2012 by A Nonny Moose  

[Mark Waybill 242]

Culinia - "apparently this is what mustard is made from: seeds, water, salt, lemon juice, mayonnaise so imagine having separate code for each of that and how that is made and the ingredients in that (like whatever makes mayonnaise, farming coding for seeds etc etc), lol. Possibilities are endless."

This is the logic behind a theory I've had since I first got Game Maker in 2005 and started experimenting with object based game creation. The idea of 'agents' and 'resources' is similar to 'objects' and 'global variables' and I, after seeing Game Maker's strengths and weaknesses, predicted that if a truly progressive city simulator was to be made, it could be done by creating an engine with a similarly simple system. Extremely or apparently extremely complex processes are all the result of simple interactions between simple objects, that's how physics works, and this recalls my first time playing SimCity 2000 when I realised how UNcomplex cities really are, whereas before I'd thought of them as mind bogglingly complex entities, I realised that behind all that jazz were simple laws governing simple 'agents' and 'resources'.

It is indeed a wonderful thing to see such a system put into practice, very progressive. And like Culinia said, as everything is relative, with such a simple but powerful tool, the possibilities are endless. By adding or removing layers of depth of complexity one can increase or decrease realism to the point where in theory if computers were powerful enough and programmers paid enough we could simulate anything. Thus, if we actually wanted to, we could of course make mustard from seeds, seeds from various plant cells etc and ad infinitum. Thankfully they're also trying to retain the fun vibe of sc2000 or we might begin to take all this seriously.

[Ocram's Razr 964]

The simplest route is bottle factory + farm --> mustard factory, though the farms could be further broken down to poultry farm, mustard farm, and citrus plantation.

Now, if we get a wide variety of farms, we could have certain things accepting any type of food (food processor, restaurant), to have a specific category of food (such as vegetables, fruits, grains, meats, eggs, dairy, etc), a plantation crop not used for food (such as indigo and cotton from civilization). The type of farm that is built (and the type of crop grown) would naturally go with supply, demand, and market value of crop. Let's say your farms are profit driven and not run by any businesses in town. The local demand for grain is greater than supply but market prices for indigo is far higher than grains (and your farms are suitable for indigo) because everyone grows grain. Your farms would grow indigo if set to global market profit driven or grain if local supply and demand driven. When running a surplus in all foods, your new farm zones will probably develop into cash crop farms.

  Edited March 13, 2012 by OcramSeattle  

[Jahnri 21]

I wonder how much demand on a global scale will play a factor.

The whole economic cycle is fantastically simple really. Easy to understand for pretty much anyone. Yet it allows for cities to grow in unique ways.

[Mark Waybill 242]

I'd love to see different crop specialisations and their economic effects, which of course would become part of the environmental aspect of the simulation. And I think if anyone can reach the ideal comprpomise between gameplay enjoyability and realism and depth it is Maxis. The idea of what you see is what you get- of Glassbox only showing you and letting you hear what is actually happening in your city, is brilliant imo. I'll be surprised if the game can be released by the time they say it will. If I was in charge of such a project I'd want at least three years hard work on it. But I'm sure their sense of priorities will be well managed. And of course if this generates enough enthusiasm it is unlikely to be a one trick pony of a rennaisance.

Food could also have cultural impact. The digital deluxe edition's national flavour packs are something I predicted years ago, and foods of course are very indicative of local conditions and tastes. In future it would be interesting to have various types of food industry based upon local demand. Another aspect to examine would be tourism, for instance in CXL they have food industry lots, and wine growing regions could generate tourist profits. Ecotourism is another interesting area for development potential- If natural features could somehow generate enthusiasm, then a city with towering cliffs or a coastline might benefit more than Flatville USA.

The potential of this engine seems incredible.

[Ocram's Razr 964]

SimCity 4 had grain, cabbage, and fruit trees. Cities XL had grain, vegetables, fruit, livestock (with the livestock not displayed properly though in the game files), and market gardens (herbs?). So there will be some variety in the farm fields and the developers say that Glassbox is a complex "see what you sim" resource (and traffic) engine as well as having mustard as an example resource chain so I would assume that we would different properties for at least 2 of the following: grain, vegetables, fruits, livestock, seeds/nuts/oil source, and/or herbs. At the very least a distinction between healthy fruits and vegetables from staple crops. Preferably it would be broken down into sour/savory/leafy/green herbaceous vegetables, sweet herbaceous fruits (cheap when fresh but mostly for processing into juice, jam, and "nectar"), fruit trees, tree nuts, starchy produce (tubers and cereals), vegetable oil sources (corn, soya, canola), meat sources, dairy, herbs and spices (flavor), fibers and dyes (for textile factories). I highly doubt that tobacco would make it on the list of cash crops because it depletes the soil heavily (requires crop rotation and fertilizer) and its only purpose is for a drug. I doubt that coffee and chocolate would make the list because they are strictly tropical plants --- unless of course, you have different climates instead of just uniform California climate of SimCity 4 and can distinguish between cold climates (snows in winter), rainy temperate (water renews much faster, rivers more common, water tables high, like the Pacific Northwest), Mediterranean climate (warm, rainy and semi-arid), inland basin/fertile desert/dry plains (most water comes from rivers but soil nutrition is high, optimal for farms), desert (hard mode, virtually no water or soil nutrition), subtropical (rainy and warm, fertile, floods are more common here), and tropical (rainy and hot, fertile but overgrown lush verdant jungles would best be left mostly alone, natural disasters are more common here).

  Edited March 13, 2012 by OcramSeattle  

[croxis 150]

The problem with too much diversity of a resource type is that it increases the instability of the economy and decreases the player's ability to understand the situation. The opposite is also problematic in that in makes a very uninteresting simulation. Resource games, be it Anno, Xseries, or Caesar, limited the resource types wasn't because of computational power but for gameplay reasons. Caesar III is probably the closest to SimCity2013's goals and it was still a fairly deep game for having about 20 resources

[Ocram's Razr 964]

The crops I mentioned are still less than the diversity in Anno 2070. All you need crop wise are fruits, vegetables, carbohydrates (grains, potatoes, sugar beers, sugar cane), fats, proteins (which could subdivide into meat, dairy, and legumes/meat substitute), and possibly spices/herbs/flavor and vertical farms with aquaculture (though unpolluted coastal cities are better off without).

Industrially, you could not only grow wood but also fiber crops and dye crops for textiles.

  Edited March 13, 2012 by OcramSeattle  

[Mark Waybill 242]

So in theory one could have a mostly agricultural region with good water flow and create a textile based industrial town utilising the water flow for energy. Especially if there was a canal or river to the sea. It is good to know we may one day be able to create cities in such a realistic industry based fashion, reminiscent of the 2000 industrial sector graph

[Ocram's Razr 964]

In theory, yes. The trailer shows a city with a water mill. Mustard was an example resource chain. Farms will make food in SimCity and GlassBox allows for a lot of diversity. You could in theory lay agriculture zones that will grow whatever is wanted most and it might be possible to build specific type of farms if RCI plopables are included (zones are included and plopable public and landmark buildings are included)

[A Nonny Moose 8,262]

Time to think outside the "box", I think. What kind of facility will be available to edit/create regions? Will user-created regions even be possible? This is one tool that needs to be delivered with the system, and it should be simple enough for a non-computer professional to use but capable of being also used at higher levels.

SC4 certainly has levels of geekism that are probably unacceptable in today's market. Even the Lot Editor is unnecessarily complicated by the nomenclature of the various elements and objects the compose a lot. The BAT, requiring a full-scale 3D editor is much worse for the novice. People these days simply have not grown up with even GW BASIC, let alone fancy "exemplars" and so on.

Neat XML-like languages are all very well for those of us who can roll with the blow, but Joe Gameplayer isn't going to want to do too many handsprings. His education doesn't prepare him for it, more's the pity.

[Mark Waybill 242]

The only programming language I ever learnt to use was GML, although I did try to learn C but it just seemed so much hard work to achieve what I could with a few clicks in GM. Also when all you have is a long script file it is way more difficult to find your way around, whereas in Game Maker it's all stored in folders and accessible by clicking on a folder/item. What I liked with GM was that you could do simple/obvious actions and commands and then use script to define them more accurately or enhance them. But of course it wasn't designed for 3d games and Virtucity never did go 3d. Might as well mention in this thread, quite a fitting spot really, that VC is now officially dead.

I doubt SC will lack user created regions. If we don't have region wide terraforming I will be very dissapointed.

[A Nonny Moose 8,262]

Very High Level Languages (VHLLs) are all very well. However there comes a time when it is necessary to get down on the iron and scratch. The underlying libraries and interpreters for GML do not allow for full optimization even though it produces "instant" gratification. This is one of the reasons that, while I like VHLLs for quick prototyping, I still revert to a language that generates code directly and can even include machine instructions or has a machine interface to critical items.

Perhaps it is because I learned programming when even assemblers didn't exist. My first programs were toggled into the machine then output on to paper tape for reloadability. Everything was done at the console. Years later, one of the mainframes I used actually was bootloaded by entering a couple of read card instructions into an I/O controller to read the 80-column card boot loader into memory and transfer control to it. And even after that, I encountered a machine that was programmed by toggling programs into the console until the Process Assembly program could be read from the paper tape reader.

Now, some mainframes I am aware of are actually bootloaded by PC's serving as part of the console. And of course the bootloader for the PC is built into the BIOS chips. The question is, my friend, is do you know what the BIOS chip does to start up the PC, and how it got on the chip in the first place? Someone has to be aware of these fundamentals.

What would you do with a naked PC with no BIOS chip?

[Ocram's Razr 964]

I know XML in general and XHTML. I forgot everything other than making a valid, well-formed XML document and basic XHTML/CSS web pages. I learned enough C++ to make custom content for OpenTower v2 (I actually contributed roughly 1/10 the code and 9/10 the art, I could never get OpenTower v1 working because it used depreciated SFML coding) and Highrise Developer (my art was downscaled back to SimTower original size with the spaces between floors being part of the rooms, which had vents and stuff, I was not really needed). I would need to start from the bottom and switch majors.

  Edited March 13, 2012 by OcramSeattle  

[Jahnri 21]

I envy you Nonny. That is so many generations before I started working with computers. Must be a interesting perspective going through all the developments. Personally I never got to far into programming. I more or less focused on case design and gaming pc designs as of late.

  Edited March 14, 2012 by Jahnri  

[Craig-Abcvs 1,595]

For the vast majority of end users including myself, the most important thing regarding the games engine is that it is robust and that it works.

That, put very simply, is all we need to know.

[A Nonny Moose 8,262]

For the vast majority of end users including myself, the most important thing regarding the games engine is that it is robust and that it works.

That, put very simply, is all we need to know.

Completely true. That is why I would not object to a delay of delivery if the reason given was that the QA cycle was incomplete. Releasing a game with bugs in it is reprehensible. A game or any app for that matter should be like a refrigerator. Plug it in, set the controls and don't worry about it again for a very long time.

Programs that crash are cases of bad programming. No program should ever ignore signals because the programmer either didn't know about it (not well educated) or the programmer didn't know what to do with it (incomplete specification). Programs must run without surprises other than those programmed into them.

For this reason, I am also in favor of an on-line environment for developing user custom content. This must be adjudicated by the server end and certified free of problems before they can be added to the repository. This may become a huge task, but if the tools provided are good, there should be little to be done at the certification end. Perhaps qualified custom content providers need to be certified by the central site?

John C. Winterton, CCP

Professor of Computer Science (retired)