Wednesday, April 23, 2008

Vault OS : Feedback Needed

I have two things to inform you of and I need to know if anyone is interested.

First, I've got Civil Defense Commander, which is say 75% complete. I've got all the source and binaries for Borland C++ Builder 4 which I could package up and release as public domain code. It makes me wince to look at because there is little room for expansion and not much attention to some of the final details of the implementation. I left off at this point because I gathered it was not what I wanted. I had to think about it for some time before I got around to the Vault OS concept. Some of you from the first Vault-Co site may remember when I had photos of that posted up. It's no skin off my nose to give it away, it might be useful to some individuals and it even might get polished up if I include all the source code. I can tell you a specific path to follow to complete it, even where you can get a free report design tool to plug in to generate printed reports. I just feel like I have other things that are better directions to go in, for example Vault OS.

Second, I've got CDC.NET, which is an application I've never told anybody about. Basically, it is a .NET/ASP.NET powered shell for a complete client/server shelter management system, running under Windows with a server required (MSSQL/MySQL). It's all powered by Web services over the network and uses stateless sessions for everything. I have not looked at it in 2+ years, but I distinctly remember it being fairly complete and quite usable. This includes inventory, logs, time management and scheduling, personnel and medical records. I could have kept working on this and was pretty close to making a complete product before I abruptly got sick of it, especially when I was thinking it was pretty current tech if it would have to also run on ancient x86 hardware. I didn't want to ever mention it since it would appear to be just something else I got started on and never finished with gusto. * UPDATE * : I just had a look at it. Much more advanced than I remember. Looks like a perfectly usable complete app, in fact. You wouldn't know otherwise unless I told you from looking at the front of it.

So there are two options here. Hendo, a friend of mine for a while, the guy who always promises to run me over with a monster truck, has recommended I do a .NET version of Vault OS first that will run on modern hardware, then do the x86 version later, because modern hardware is much easier to configure and develop for.

1. Release both CD Commander and CDC.NET as public domain or open source. Kiss'em goodbye and never look at 'em again. Stick to working on Vault OS in QNX Neutrino for x86 PCs in DOS.

2. Release CD Commander as Open Source, then do the work needed to package up CDC.NET and release that as a test version of "Vault OS for Windows."

I'm curious to know what would be more useful to the most people in the most accessible way in terms of platform, environment and paradigm. Remember CDC.NET will work standalone on any machine with the server running. There's also another angle ... CDC.NET may be very demanding (relatively) for the server computer (you'd need a modern portable with Windows 98SE/XP at a minimum) but the clients can be pretty much anything running a browser. So you could still use cheapo PCs even running something like Arachne possibly (?) under DOS for the client end, as long as the server was a decent machine. I know it can be done and is routine for many sites but I don't know the exact mechanics of plugging DOS boxes into area networks in Windows. Of course, you could also just run the client browser on another low end portable with Windows of some flavor. I think the problem with pure DOS browsers is that CDC.NET uses Javascript for some client-side validations. That still doesn't mean it wouldn't work.

There's one more interesting twist to add to this. You can run a real .NET application on a Windows platform with an embedded HTTP browser as the client. Above this browser window you could put any kind of Winforms/Toolbar/Tabbed/Menu you wanted to launch things like a local monitor for I2C sensors like I was talking about for Vault OS in DOS. There's also the potential there to submit local sensors to the server in the background so they are available to all other machines on the network.

Finally, considering what I already have running, this is much easier to develop than my grand plans for Vault OS in QNX Neutrino. It's coming along well but of course you're still talking about ten hours of coding in QNX to duplicate functionality I can perfect in ten seconds by dragging it from the Visual Studio IDE and dropping it on the design window.

Any feedback much appreciated. If it's not obvious by now, I always have more ideas than time.


Anonymous said...

make it a windows based system, gui interface, keep in mind many people are not IT savvy and whilst comfortable working within the windows environment will be outside their comfort zones with your other options. SQL servers are pretty easy to set up and run. MS products have service packs, updates etc. I have seen some real nice databases using .net that are only limited by your imagination. Reading your post, you are all over it and have the answers already. Good luck by memory the screen shots showed real promise.

Texas Arcane said...

Thanks a lot for that feedback. I am honestly going to allow people to set my direction on this because if there is one thing I have always been terrible at it is guessing at what others actually need in terms of technology.

So if there is similar feedback that people never really considered acquiring a windows platform to be a problem, that is a very rich direction to pursue for obvious reasons. It doesn't mean that later on I can't construct clients to run in QNX DOS.

Chesterton said...

Some things to consider:

An automated assistant is going to be most useful during the first few weeks of WWIII. Have you ever had a dream so real that you felt like you'd actually lived and experienced it even after you woke up? I had one of those a few nights ago about the start of WWIII. I felt blind terror. Despite having very detailed plans for a TSHTF situation -- the best I can do for my current situation -- I still felt confused and somewhat lost. I think it will be like that for just about everyone, regardless of how prepared they are. So, an easy to use manager which can help you get organized quickly would be a great help. Something you could just boot up and see:

- What is going on in your shelter and what you need to do to make sure everything is working well, and that you aren't going to die because of something stupid like closed air vents or something of the sort. It sounds like an idiotic way to die now, but I think most shelter deaths occur because of these small things. People panic, people forget some small detail of shelter operation, people die. Having a simple automated guide would cut down on such deaths.

- Details about your food supplies, daily tasks, etc. Again, this will help people get into the rhythm of shelter life right away, while they're still stressed. A list of jobs would probably ease the stress levels.

Since this will only be "necessary" during the first few weeks, before everyone has familiarized themselves with every inch of the shelter, a Windows machine would do just fine.

A few comments about Vault OS. While I think it's a great idea to make it run on old, reliable hardware that can go on working for much longer than your typical Windows system, I think planning ahead that far with computer hardware might be a waste of time. There are so many things that can go wrong in 10-20 years, especially in a post SHTF world, that I would be incredibly surprised if the hardware was still running after that time. Humidity, moisture, flooding. Deep cycle batteries have a shelf life and do fail. Solar panels fail. Wind turbines break and are more difficult to fix than everyone believes. I think I would be more than satisfied with a Windows based machine running a simple application that helps me monitor my shelter and get organized during the most stressful time of the war. Ease of use is key.

If the machine fails after a couple of years, oh well. There'll be plenty of time at that point to find another way of tracking everything. After several months of living in a shelter all of the inhabitants will know it like the back of their own hands, and the only things to worry about at that point are thieves and cobalt. It will not be difficult to adjust to life without a computer then.

I should mention that I don't currently have a shelter and won't for some time, so all of this is just a theoretical exercise about what I would want. Theoretically.

Texas Arcane said...

Very useful. Thanks for that.

Anonymous said...

I do have a shelter, although I consider it bare minimum and am working my way towards one of those nice 'faraday' steel culvert setups like you have.

I agree, an automated system is absolutely critical for the first phase of shelter dwelling. You never know how preoccupied you'll be with medical treatments, you may have been caught away from home and be facing some radiation exposure by the time you get underground.... that's when you need to be laying down and recovering, not trying to monitor all your life support systems.
There are a thousand variables of health and psychological state of you and any other occupants that can cause enough distraction that it's unrealistic, or at least too risky to be considered sound preparation, to assume that you will always have 100% calm attention on life support systems.

The things I would expect from a vault computer are simplicity. As low-tech as can be managed, and simple. Here's my wish-list that I would love to see on the main, running and open screen. (I will be sticking to green/black monochrome, BTW...just personal prefference. I'd get red if I could, to be more compatible with powered-down, low-lighting situations.)

First and foremost, a simple clock. Being able to easily see the passage of time is very helpful to people stuck in these isolated situations. A guy can go nuts staring at the wall for an unknown number of hours or days.

Feeds from Two internal and one external. I would hope it wouldn't be too difficult to plug in a mathematical function to monitor the rate of reduction of external radiation, and produce an estimated readout in months, days, hours until the exterior is safe. This might prove vital in resource management for some, and would be a comfort to most people along the same lines as a date/time clock readout.

Most critically, I need it to manage the air quality... this can be as complex as O2 sensors, CO2, humidity, temperature etc.... or as simple as a variable timer for the circulation system.

Things like water level, internal temperature (not external though, which is a minus) power remaining, power consumption, power production, these can all be monitores with local gauges and instruments. It would be really nice to have these things hooked up to the system, but not vital.

I disagree that a windows-based system is that critical to most people due to familiarity. I'm unfamiliar with really ANY programming terminology, but I've used many different types of simple programs and i don't think anyone would have a problem once familiar, if the system is simple enough.

One main screen with the pertinent readouts and info. use the arrows to highlight different ones. hit enter to open a page where further info on that element can be seen, or adjustments can be made, like the time intervals and duration of the circulation system, or change the clock. Simple. Like the BIOS menu in my computer at home. A skin would be awesome, I'd get a big kick out of a 'pip-boy' type of skin, but I don't think anyone will complain about a simple, functional utilitarian readout screen.

I also have no problem with you tailoring the system to be compatible with only certain products. I think enough people would prefer a simple 'plug and play' fully supported and compatible system, that they'd be willing to buy the input components this system is programmed for, especially since we all know you're good at finding the cheapest gadgets. Leaving the inputs open for each person to 'customise' and program, Could leave many laypersons stumped. like myself.

My $.02.... and at today's exchange rates I hope that doesn't put me in debt. hehe..

(A Thousand Good Intentions)