Re: [Gnumed-devel] re: vpn justification for xml-rpc

[Tim Churches]

On Tue, 2004-06-08 at 23:56, Ian Haywood wrote:
> On Wed, 09 Jun 2004 19:29:44 +1000
> sjtan <address@hidden> wrote:
> 
> > Or writing a separate web client.
> IMHO web clients are unusably slow even on the fastest network as a design 
> issue
> Remotely, it would just be too painful. 

Not really. Properly written Web apps can be very fast - certainly not
perceptibly slower than a local GUI client - provided that the latency
and bandwidth to the server are low enough. We have been testing a
Web-based public health app and when accessed via the Internet, even via
a modem, from capital cities around Australia, new forms appear almost
instantly. When the server is on the LAN or WAN, it is absolutely
instantaneous.

The real problem with a Web interface is the impoverishment of the user
interface and the lack of interface widgets in HTML. To fix that, you
need to start using JavaScript, which means a nightmare of debugging on
multiple versions of multiple browsers as soon as it starts to get a bit
complex, or various borwser-specific plug-ins eg ActiveX on IE, XUL
components on Mozilla etc). Or you can write your client in Java, but
then you might as well have written with Python and wxPython...

> A text-only client running via SSH would be interesting, Sebastian has 
> floated this previously.
> 
> > If HL7 v3 becomes the common wire protocol for medical record messaging,  is
> > there any need for xml-rpc?
> AFAIK That's a long way into the future. The HL7 guys also still aren't very 
> FOSS friendly:
> you need big $$$ to see the specs.

No, you just need to join HL7 Australia, which is a hundred bucks or so,
I think (dunno, we have a corporate membership).
-- 

Tim C

PGP/GnuPG Key 1024D/EAF993D0 available from keyservers everywhere
or at http://members.optushome.com.au/tchur/pubkey.asc
Key fingerprint = 8C22 BF76 33BA B3B5 1D5B  EB37 7891 46A9 EAF9 93D0