[Info-vax] OpenVMS Development Annoyances

[Dave Froble]

On 4/21/2019 5:51 PM, seasoned_geek wrote:
> On Sunday, April 21, 2019 at 10:49:52 AM UTC-5, Dave Froble wrote:
>> On 4/21/2019 11:21 AM, seasoned_geek wrote:
>>> On Sunday, February 10, 2019 at 1:36:17 PM UTC-6, Dave Froble wrote:
>>>>
>>>> And what if that one year in 100 is next year, or this year?
>>>>
>>>> But, let's make it a bit more personal.
>>>>
>>>> Let's define a circumstance, perhaps a bit far out today, but perhaps
>>>> not so in the future.
>>>>
>>>> You have a medical problem, and the only treatment is computer
>>>> controlled surgery.  If things do not go flawlessly, for whatever
>>>> reason, then you die.
>>>>
>>>> So, will you accept that computer control from some cloud, or perhaps a
>>>> local dedicated system, or perhaps a disaster tolerant cluster where a
>>>> dozen systems must fail to fail the surgery.
>>>>
>>>> So, Ok, your life, you choose, and I won't even make any suggestions,
>>>> nor fault you on an improper choice.
>>>>
>>>
>>> Speaking as someone who has been working in the medical device world, at least in America, the FDA doesn't allow cloud hosted surgical robots. You generally aren't even allowed any inbound connection. If the device has any network access it must reach out to a hard configured address then either push or pull data in a provable risk tested manner.
>>>
>>> If a device has any risk to human life short of electrical surge which has to be mitigated via hardware, you can't even use Linux in that portion. That's why you see those fanny pack chemo devices with multiple processors in them. Life risking functions are on one or more processors running an approved RTOS. Network communication of any kind has to have its own processor and RTOS.
>>>
>>> The processor running a user interface, be it touch screen or not, can run Linux if and only if the hardware design of the device physically blocks access to network.That became mandatory once the Linux kernel rolled part of networking into itself.
>>>
>>> Having said all of that, yes, there are genetic misfits out there attempting to circumvent "all of those burdensome regulations." Many/most of them are career academics.  One in particular is a Phd from Johns Hopkins trying to run a post-op clot monitor on an ordinary laptop running windows. If they somehow manage to bribe that through FDA approval, it will be on those late night 800-number lawyer commercials within weeks of shipping.
>>>
>>
>> Interesting info.  But, on that last point, I'd suggest that there is a
>> bit of a difference between "monitoring" and "doing".  For example, I'd
>> accept a lesser environment for monitoring blood sugar levels, but might
>> want a bit more from something administering insulin.  Regardless,
>> better safe than sorry.
>>
>> --
>> David Froble                       Tel: 724-529-0450
>> Dave Froble Enterprises, Inc.      E-Mail: davef at tsoft-inc.com
>> DFE Ultralights, Inc.
>> 170 Grimplin Road
>> Vanderbilt, PA  15486
>
> There's not. At some level the monitoring stuff gets hooked to the doing stuff. How do you think a surgical robot knows an artery sprung a leak? Drop in blood pressure? Etc. etc.
>
> They don't invent a shiny new monitor for the robot. They take a bunch of stuff which already exists __AND__ has gotten FDA approval.
>

Is that anything like FAA approval for the Max 8 ???

-- 
David Froble                       Tel: 724-529-0450
Dave Froble Enterprises, Inc.      E-Mail: davef at tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA  15486