@ Charles (and anyone else)
using
tsNetSetTimeouts 30, 0, 300000, 60000, 30, 1000
I need to implement this in our app… "minimum amount of data…" is obviously
different from "no connection whatsoever"
Your example assumes that for this use case, 30 seconds is acceptable. That may
well be for a DropBox app where "downloading stuff" is a known "may take a long
time" context. But clearly that's way too long for many contexts. Mobile users
have high expectations on responsiveness. Can anyone give us some standard
numbers for a Bytes/Seconds setting(s) for a reasonable time out period for
e.g. streaming audio OR an HTML page?
I realize this question is highly context dependent. If one is calling a plain
text static html page with one small image, the total data is very little. even
on a slow connection, it might still load in a "some" (reasonable wait)
seconds" … but then if one is about to stream an audio file (mp3) from the
server which could be e.g. 30 minutes long in duration, the player will wait
and wait and wait perhaps get started, then stall ….since we set up Cloud
Flare I'm seeing some big improvements on this front. no more outgoing
traffic/CPU overload notifications from our Linode server…since CloudFlare is
now serving all cached files from their CDN… Bu we still need to deal with
local latencies as the user moves from her Home to his car to Starbucks + WiFi
and then a walk down the street into the office (back up on hi-strength wifi
Spotify is quite "brutal" in it's approach… with 2 bars on a 3G connection… I
can still go to Safari in my iPhone and fetch a small html page, but Spotify
throws up it's ugly disconnected icon with no information on a black screen.
Flipboard (delivers articles from publications… less bandwidth needed that
Spotify for audio) used to just not open at all… now it will give some message
about "connect to the internet" -- confusing to the user when he can still
connect in Safari
So if that's the way the Big Boys with the Million Dollar Apps do it. I wonder
if I am just dreaming that us "little LC devs" can do a better job.
With TSNet (thanks for making this fully available now in Indy!) we could fine
tune this along the way. Theoretically one could dymanically reset the time out
ahead of different classes of content
1) skinny text only pages
2) image heavy pages
3) streaming audio/video.
So I guess this boils down to one question:
If you are streaming audio or video.. .what would be the reasonable bytes per
second you would want to "see" as current bandwidth before telling the users
that their connection is too slow.
All thoughts on this welcome, as I go off to the net to find bandwidth
standards… But maybe someone here already knows the answers/has experience? And
of course, I can just build it and test (which we will do anyway)
Brahmanathaswami
On 11/22/17, 6:16 PM, "use-livecode on behalf of Charles Warwick via
use-livecode" <use-livecode-boun...@lists.runrev.com on behalf of
use-livecode@lists.runrev.com> wrote:
Have a look at the following bug report which is about a similar issue:
http://quality.livecode.com/show_bug.cgi?id=20627
Effectively the last two parameters to that command allow you to set a
minimum amount of data that must be transferred across a connection within a
specified period of time to consider a connection still active.
Having said that, when tsNet is disabled, libUrl should continue to pay
attention to the socketTimeoutInterval property.
I have just done quick test on LC8.2.0dp2 and if I set the
socketTimeoutInterval before making a “put URL xxx” call to a script that
deliberately doesn’t respond in time, the request times out after the time
specified by the socketTimeoutInterval call and returns a result of “socket
timeout”.
_______________________________________________
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode
