Communicating between to standalones with "write to process" - Found word(s) list error in the Text body
dochawk at gmail.com
Tue Dec 3 03:31:44 EST 2013
On Sun, Dec 1, 2013 at 7:22 PM, Mark Wieder <mwieder at ahsoftware.net> wrote:
> Half a second is quite a long delay. You could bounce signals off the
> moon in that time.
Yes, but not across Las Vegas . . . :)
At home, I'm on a Clear wireless connection. I have a bunch of blocks of
put the milliseconds into strt
put the milliseconds - strt into stp
stp is generally 500 ms, give or take.
When I"m actually on the localnet, whether on the server itself or my
laptop, I don't notice a lag.
A threaded stack would solve this for me, but as it stands, the database
functions take a long time, so I have to hide them from the user.
Primarily, I watch time since the last keystroke, and if it gets to a
couple of seconds, hitting the db is fair game.. But for my needs, a
semaphore system would be sufficient: People *could* enter conflicting
things on different systems,but for this type of application, if that is
happening, there is a bigger problem in the firm than I can hope to help
with (It means that Jane is entering data which conflicts with with what
sue is entering. You either owe Macy's$500 or $700 . . .). As such,
updating the local db to just have the information that soeone else entered
is enough, even if it takes a couple/several seconds. Blocking the user,
though, is *bad*.
(OK, I'm an impatient person, and my software reflects this. 10+ years
ago, I had an SE/30 rather than an SE over this . . . Never buy your
software from someone who isn't lazy & impatient!)
Dr. Richard E. Hawkins, Esq.
More information about the Use-livecode