W3C home > Mailing lists > Public > www-svg@w3.org > February 2013

Re: High Resolution Time

From: David Woolley <forums@david-woolley.me.uk>
Date: Sat, 09 Feb 2013 15:46:47 +0000
Message-ID: <51166F67.3070207@david-woolley.me.uk>
To: www-svg@w3.org
Jonathan Chetwynd wrote:
> "accurate to a thousandth of a millisecond."

This is not about accuracy, only resolution and monotonicity, see below.

> http://www.w3.org/TR/hr-time/
> 
> notwithstanding the moniker,
> has anyone an insight into adoption and use of accurate microsecond 
> timescales within SVG?

One microsecond time accuracy is not easy.  It generally requires a 
local GPS timing receiver (not the more common navigation ones) and 
careful attention to network design.  It also generally requires the 
right operating system (Windows is not particularly friendly to high 
accuracy timing).  In particular, w32time, out of the box, is not 
suitable for such accuracies, and even fully tuned, you really need 
something better.

The average PC user, these days can expect time to about 100ms.

With network time sources. and some active management, 1ms is reasonably 
achievable.  1 microsecond is achievable, with a GPS timing source, in 
PPS mode, with Linux or BSD Unix systems, and suitable, low latency 
hardware.  0.1 microseconds is getting close to the limits of GPS.

> 
> there's a clear use case for scale as in time,

What is that obvious use case?

> and perhaps an expectation regarding HTML5 visualisations that may also 
> map canvas.
> 

Time intervals are somewhat easier, but you need good time 
synchronisation software to better 10ppm and you will probably need GPS 
timing to get better than 0.1ppm.

The article seems to be about relative times measured over short 
intervals (i.e. not with extremely low percentage errors).  It is about 
high resolution, not high accuracy, time.

-- 
David Woolley
Emails are not formal business letters, whatever businesses may want.
RFC1855 says there should be an address here, but, in a world of spam,
that is no longer good advice, as archive address hiding may not work.
Received on Saturday, 9 February 2013 15:47:22 GMT

This archive was generated by hypermail 2.3.1 : Friday, 8 March 2013 15:54:54 GMT