Hi,
I'm absolutely new to this list but I just want to throw my thoughts in.
On the bigbrother extensions-site I saw a perl script implementing the
bigbrother protocol.
Taking this script as a starting point, using some perl-modules from
the Win32::-family, making use of the WMI (
http://msdn.microsoft.com/library/en-us/wmisdk/wmi/wmi_reference.asp )
and a bit of hacking would end up with a hobbit-windows client.
In our IT department (where we use many windows servers,
unfortunately), it is not applicable to have cygwin on the servers. So
the best way would rather be writing a hobbit client either in perl or
C#.
....
Sorry for mentioning C# and discrediting cygwin, but I wanted to state
that (IMHO) cygwin is not the (business concerning) answer...
regards,
Manuel
----- Message from user-ce4a2c883f75@xymon.invalid ---------
Date: Mon, 28 Nov 2005 10:25:48 +0100
From: Henrik Stoerner <user-ce4a2c883f75@xymon.invalid>
Reply-To: user-ae9b8668bcde@xymon.invalid
Subject: Re: [hobbit] Hobbit client for Windows
To: user-ae9b8668bcde@xymon.invalid
On Mon, Nov 28, 2005 at 04:44:30PM +0800, Vernon Everett wrote:
The biggest problem with a Windoze client will be the inconsistency of
Windoze.
With some versions, a command is standard. With others you require a
service pack. Still others require the resource kit. Some versions of
Windoze just don't have some commands. It gets quite ugly.
I have been informed by my Windoze admin colleague that the behaviour of
commands can differ between versions too.
Cygwin is consistent, but not complete.
Indeed - despite all the talk about Unix being "fragmented", the
Unix API and command-interface is immensely more consistent than
the Windows ditto.
I now understand why Henrik hasn't produced a Windoze client yet.
It took me about a week to write all of the Unix clients - Linux,
*BSD, AIX, HP-UX, plus the server-side back-end - after which I had
a rough implementation that mostly worked, but needed some testing.
It's been about 7 years since I last had the misfortune to do any
programming on Windows, so writing a Win32 client would probably
take significantly longer than that. The Windows "performance counter"
API is just ... abominable.
(If you want to see some live examples of the different ways of
thinking in the OpenSource world vs. the Windows world, have a
look at this comparison between the Open Document (.odt) XML format
and Microsoft's .docx XML format.
http://www.groklaw.net/article.php?story=20051125144611543
If you thought using XML would help with interoperability,
you're in for a surprise).
My biggest concern for a Windows client is how to support it.
Regards,
Henrik
----- End message from user-ce4a2c883f75@xymon.invalid -----