CT Scout ship computer ???

[Travellerspud]

as I realized earlier;

a ship can have essentially an infinite number of programs that are not stored in the computer (like a collection of old Atari Game cartridges on your bedroom floor)

The 2 capacity numbers define the total size of programs that can be stored/loaded in the computer. The computer automatically switches these in and out of storage as-needed each phase (as selected by the operator).

The end-of-Turn re-programming interphase is where programs are removed from the computer (dropped on the floor, so to speak), and other programs loaded into computer storage (from the floor)...

 

[Brandon C]

Remember that only 1 MCr of programs is free for the ship, as well. The list above looks like it would add 20-30% to the cost of the scout.

 

[kilemall]

But not both loaded into the computer at the same time. I think perhaps you two are talking past each other.


Hans

I have clearly stated the situation, I shall leave it to him to open up the books and reread. If the starter set says so and he's happy with that, fine, IMTU choice and all, but several people telling him the same thing should be a trigger to at least check.

 

[kilemall]

Remember that only 1 MCr of programs is free for the ship, as well. The list above looks like it would add 20-30% to the cost of the scout.

Of course, fighting programs always rack up the charges under the LBB2 regimen.

Which is why the computer programming part of the rules are so enticing, and so dangerous with the bugs.

 

[kilemall]

as I realized earlier;

a ship can have essentially an infinite number of programs that are not stored in the computer (like a collection of old Atari Game cartridges on your bedroom floor)

The 2 capacity numbers define the total size of programs that can be stored/loaded in the computer. The computer automatically switches these in and out of storage as-needed each phase (as selected by the operator).

The end-of-Turn re-programming interphase is where programs are removed from the computer (dropped on the floor, so to speak), and other programs loaded into computer storage (from the floor)...

It's 4 for the Model1/bis and no phase switching except during reprogramming.

And best keep your floor clean and your boards racked, would be a shame to step on a 10 Mcr program and crack it, literally.

 

[Brandon C]

But not both loaded into the computer at the same time. I think perhaps you two are talking past each other.

No, not at the same time.

 

[kilemall]

No, not at the same time.

What, you are saying you need the program available, just not 'slotted'?

Of course you do, Jump-2 does not confer the ability to do Jump-1 inclusively, you have to have both to have the choice.

That isn't a factor in your critique of the Mod/1 Bis being unusable for the ship, my example is an extreme since Jump-2 requires more space. Jump-1 requires less space and therefore is less constraining on computer resources.

Example, not every configuration for every jump type.

Travellerspud was listing all these programs, we corrected his understanding but it sounded like you were arguing the case that they were both needed loaded at jump time.

 

[Brandon C]

What, you are saying you need the program available, just not 'slotted'?.

Yes, that's what I'm saying. It was left out of the list of all the programs a scout might need.

 

[Ulsyus]

Nope: for a ship off the shelf? Sure, that looks about right.

An active duty scout, or any ship that's been in service for any length of time, would have a more extensive library for sure.

Sensor data and the like is sort of handwaved: IMTU I assume it to be a background feature of the maneuver and navigation programs.

Were there even sensor rules included in starship combat in CT, or is it just assumed that ships can just pick each other up out of any clutter, and that any modifiers are based on the gunnery, missile, and ECM programs/rules?

 

[rancke]

And best keep your floor clean and your boards racked, would be a shame to step on a 10 Mcr program and crack it, literally.

I don't believe in programs costing that much. Considering that programmers cost perhaps 50,000 credits per man-year, you might have to pay millions for developing a really complex program, but if you are selling to thousands of customers, the individual programs would cost anywhere near that much.


Hans

 

[kilemall]

I don't believe in programs costing that much. Considering that programmers cost perhaps 50,000 credits per man-year, you might have to pay millions for developing a really complex program, but if you are selling to thousands of customers, the individual programs would cost anywhere near that much.


Hans

Certainly if you are looking at modern pricing practices for business users that's true.

However this is exceedingly expensive must not fail software, at an avionic level of utter reliability and fault tolerance beyond present known levels, which we can see with the computer hit/failure rules.

No bugs, and it takes a near nuclear hit to take it offline.

That's not commodity level software OR hardware performance.

The closest thing I can think of performing to this level is the Saturn V computers, or the US ABM computers, one could rip cards out and mess with power supplies and they would still run- in the 60s.

But even so, yes across an installed base of millions of computers delivered over 1000s of years of human interstellar travel, a lot of this should already be already 'sunk dev costs' and a done deal.

However.

The costs are not so ridiculous when you consider how long the software is in use.

If one went more like current practice, you would purchase the license and rights to run (and in our backplane model, a really tough card) for say 200,000 Cr.

Then you would pay maintenance and support, say 20,000 Cr.

For 40 years.

That's 800,000 Cr, for a total of 1 MCr.

So, not so unreasonable looking at it that way.

Now consider what is at stake with the average starship operating culture.

Billions, even trillions of credits worth of ship at just one A starport.

As much as that is, the total value of the average midtech planet is staggering.

There is NO room for starship loss from misjump, detonating, failing in combat, or impacting on planets and space facilities due to glitchy software.

So given what is at stake, it would be reasonable to expect a VERY intense VERY costly QA effort.

And because the Imperium or whoever is going to blanch at the idea of insuring or accepting liability for planetary losses due to starship computer malfunctions, the policy would be, NEVER, and if NEVER becomes HAPPENS then the software company pays.

I'm expecting there are some huge several trillion credit bonds sitting in a vault somewhere that covers liability, and so servicing that liability coverage AND having a testing regimen with every thinkable combination of ship and electronics and drives and weapons plus whatever validation at build/modify time costs BIG BIG money.

Couple that with

* economic downturns where captains and merchant lines would cancel their support contracts and attempt to slide by while liability does not drop a bit,

* banks who would like to talk owners into putting their software on the monthly financing tab and so likely encouraging software companies to have high prices,

* some of the phenomenal gains one gets with the software that are cheaper then hardware and

* the unending nightmare of making software work with equipment that has to work with brand new manufacturers that didn't exist when the computer and software was installed alongside 30 year old equipment from designs that no longer exist

and yes I think a case can be made for the software companies wanting their money up front.

 

[rancke]

However this is exceedingly expensive must not fail software, at an avionic level of utter reliability and fault tolerance beyond present known levels, which we can see with the computer hit/failure rules.

No bugs, and it takes a near nuclear hit to take it offline.

That's not commodity level software OR hardware performance.

But neither is it a performance level that I find it credible that the average shipowner will be willing to pay for if he can get something that works almost all the time and save himself a few million credits.

Now consider what is at stake with the average starship operating culture.

Billions, even trillions of credits worth of ship at just one A starport.

As much as that is, the total value of the average midtech planet is staggering.

There is NO room for starship loss from misjump, detonating, failing in combat, or impacting on planets and space facilities due to glitchy software.

So given what is at stake, it would be reasonable to expect a VERY intense VERY costly QA effort.

And because the Imperium or whoever is going to blanch at the idea of insuring or accepting liability for planetary losses due to starship computer malfunctions, the policy would be, NEVER, and if NEVER becomes HAPPENS then the software company pays.

Now consider what other regulations would be in place if the Imperium really did embrace that level of paranoia, but are conspiciously absent. E.g. an absolute ban on jumping with unrefined fuel.

Not to mention an absolute ban on using computer programs written by PCs. ISTR rules for PCs with computer skills writing their own programs. Or am I misremembering?


Hans

 

[kilemall]

But neither is it a performance level that I find it credible that the average shipowner will be willing to pay for if he can get something that works almost all the time and save himself a few million credits.

Sure, which is why I have all those IMTU rules about computing down to 16,000 Cr computers with 3,000 Cr programs- but the ship isn't certified for High Passage and the commodity computers break pretty easily.

A whole world of semilegal tramp passenger and cargo traffic with broken down leftover ships and cheap computers- but you can go 6 parsecs on 3000 Cr.

Now consider what other regulations would be in place if the Imperium really did embrace that level of paranoia, but are conspiciously absent. E.g. an absolute ban on jumping with unrefined fuel.

Not to mention an absolute ban on using computer programs written by PCs. ISTR rules for PCs with computer skills writing their own programs. Or am I misremembering?


Hans

Not paranoia- just comparable to passenger jet certification today.

No laws against it, just certification, law of the lawsuit jungle taking down underfinanced medium-sized software companies, greed, a nice high bar against threatening competition, and encouraging private financing of public safety.

I assure you, the Beowulf's owner will get sued into oblivion if that ship plows into a highport due to their own bug- they just aren't as deep-pocketed as a software company.

 

[aramis]

Not to mention an absolute ban on using computer programs written by PCs. ISTR rules for PCs with computer skills writing their own programs. Or am I misremembering?


Hans

They're present in CT 2E; in CT 1E, the writing softwear was an addon.

 

[nobby-w]

But neither is it a performance level that I find it credible that the average shipowner will be willing to pay for if he can get something that works almost all the time and save himself a few million credits
...

Working in I.T. you get to see a lot of middle managers who have the rather naive view that they can set a timeframe and budget and get an arbitrarily complex application delivered to their specification. This mentality is the root cause of most commercial software project failures.

Engineering (including software engineering) has this rather annoying property that folks who aren't involved in the field often fail to appreciate: You have to build something that actually works.

The system isn't finished until it does the job. This can involve a lot of frigging, testing and refining - a process often known as 'ironing out the bugs'. The image of a cowboy programmer knocking up a complex system in his basement and having it work first type is a myth. A small handful of genius programmers have been able to make complex applications single-handed, but it's not how software development works in practice.

The other thing about designing software for avionics systems is that you also have to flight-test it. High-reliability safety-critical software is expensive to develop at the best of times, but this type of software will also require a significant number of flight-hours in a live spaceship to test it, let alone get it certified. Do you want to be the guinea-pig for an un-tested Jump-3 application?

Your average free-trader captain isn't going to be up for spending money on a flight testing programme. Software that works 'most of the time' isn't going to be terribly attractive if it fails at a critical moment like landing (something like this actually happened on the Apollo craft).

Say, for example, I wanted to write a Predict-5 application for a starship fire control system. How do I go about testing it - charter a few ships boats to act as targets for a dummy run? The charter company isn't going to be well-pleased if you return their craft full of laser holes after your live-firing integration tests. Didn't test it on a live test bed with the real turret and weapon system? Dollars to donuts there's a great screed of integration issues just waiting to make the system fall over and crash. Didn't you know that early models of the Mk 46 turret manufactured in Glisten before 1103 have to use the version 2.1 protocol (with proprietary extensions) unless you can get a patch from the manufacturer¹. If you use 2.2 or later they will freeze if you attempt to traverse at more than 30 degrees per second. Gee .. that wasn't in the spec.

Designing this type software is a highly specialised racket. Your average PHP monkey off the street won't even know where to start on writing a system like this.

IMNSHO people won't be inclined to monkey with the avionics software on their starships if you can buy it off-the-shelf. You could reasonably assume a secondhand market for this (with substantially reduced prices) or maybe some dodgy cowboy operators writing software in a jurisdiction where the certification authorities can be bought off. However, the latter would be an exceptional case.

But, a free trader won't have the time or skills to DIY and the actual costs of developing software like this (let alone getting it certified) will far outweigh the cost of purchasing it.

¹ By the way, the patch is classified as a munition and you need to get a licence from the naval attache at the subsector capital to obtain it.

 

[Patron Zero]

Mind a slight side-bar but wouldn't a ship needing to meet criteria to receive Imperial 'certification' to operate within it's borders be required to use 'approved' software upon-after inspection ?

 

[kilemall]

Mind a slight side-bar but wouldn't a ship needing to meet criteria to receive Imperial 'certification' to operate within it's borders be required to use 'approved' software upon-after inspection ?

That's the way it is IMTU, but that's a choice.

The CT computer rules clearly make operating homegrown programs possible (although the real cost is getting people with the requisite skill together, very tough with some of those very high skill levels, likely an adventure itself getting those kind of people to sit down for weeks of programming and testing).

The CT program cost structure is there too, and while I also found it tough sledding to justify those prices, even back in the 80s, those programs DO make the ship considerably more powerful. I guess it was easier to upgrade the ship this way rather then write extensive engineering/build rules to do it with hardware tweaks.

Anyway, I ultimately decided to not trash the entire starship building ecosystem and rules effects as they are part of the Traveller Trip.

Instead, made it a choice whether one wanted to be completely legit, operate at a merc level, or be operating at the 'edge of the system', engineering/safety, legally and astronomically.

 

[rancke]

Mind a slight side-bar but wouldn't a ship needing to meet criteria to receive Imperial 'certification' to operate within it's borders be required to use 'approved' software upon-after inspection ?

Could be. OTOH, it's canonical that ships are registered to individual worlds, so certification may be considered the domain of member worlds. It's a grey area. Perhaps claiming the right to decide whether a world's ships are allowed to fly or not was a hard sell during the infancy of the Imperium, back when Cleon was trying to entice worlds to join up voluntarily?

I could see the argument go either way.


Hans

 

[nobby-w]

Mind a slight side-bar but wouldn't a ship needing to meet criteria to receive Imperial 'certification' to operate within it's borders be required to use 'approved' software upon-after inspection ?

That would almost certainly be the case if you wanted to operate the ship in commercial shipping lanes (which would roughly translate to orbit around any inhabited world) or carry passengers or even cargo commercially.

 

[BillDowns]

I work with health care IT. On a project now for a net new install of hospital system for ~550 beds plus Level I Trauma center and several ambulatory clinics. Vendor price is around $15M.

The Vendor also operates the hardware base - several HP minis running a *nix flavor. The Vendor is also under obligation to keep the software certified for Meaningful Use, and compliant for State regulations.

So a price of 1MCr for starship jump navigation doesn't faze me at all....