3 ms·
The release of door interlock used to be as you described, but now in Europe train stop is controlled prior to releasing it. The delay you note is generally tie
by crocal 8y ago
The release of door interlock used to be as you described, but now in Europe train stop is controlled prior to releasing it. The delay you note is generally tied to the reaction time of the ATP computer in charge of authorizing doors opening. Some of these could be slow (sometimes 5s). Today it’s less than 0.5s.
Disclaimer: I design and deliver such systems.
- pjkundert 8y agoAlmost every delay in computerized systems that is greater than the physical/logical sensor debounce interval (ie. on the order of milliseconds) is due to programmer error. It plagues software literally everywhere. We need to stop! Every gas pump, payment terminal, elevator door is a testament to poor implementations of event driven state machinery... </rant>
- romwell 8y agoAnd every new tech is shinier and slower than the old generation. Nobody pays any attention to latency as a part of UX (except people making pro-photo and music gear, it seems - even gaming these days accepts pretty large latency as a part of the experience).
- crocal 8y agoI am with you in this general rant. However, in this particular case, the delay was a consequence of the introduction of computers and asynchronous telecommunications. I was not around at that time but it seems nobody anticipated that putting all the logic in a computer would result in worst performances as opposed to older designs were computations were essentially carried out by relays wired together. Relay logics worked in parallel, when CPU have to do things sequentially. Fun fact: some operators around the world refuse to use computer-based systems for some applications because of that, and they are getting a strong support now because of cyber security threats. A relay is hard to hack :D