Hi,
Nicolas Mailhot wrote:
Le dimanche 24 avril 2005 Ã 23:54 +0200, Mathias Bauer a Ãcrit :
Let's just say that linux users and sysadmins strongly disagree with you
for their own reasons, that they yelled at every single software system
that tried to do this very thing, and that none of the proponents of
this way of working ever managed to convince them it was a good way of
working,
This is probably mostly true for sysadmins. I doubt that this holds for users as well. If users can't add the browser plugin or macro package that they - and maybe only they in the entire group or organization - need or want, they will be the ones getting upset. Going through a bureaucratic helpdesk process to get this set up (or worse, rejected) is no fun. And even sysadmins may not be happy, if they can't use a needed package, because noone has packaged it in their package format of choice yet and manual install is not possible. If they need the add-on urgently - or they only want to quickly evaluate it - such a quick and dirty install is exactly what they need.
and that trying to force the issue will only result in the app-specific software updater patched out of existence on all distributions that care about oo.o.
If you remove the underlying implementation and/or CLI interface of the OOo package installer, you'll have a hard time installing any add-ons into OOo - even through 'proper' packages.
BTW: I don't think we are forcing any issue. We provide a feature, and distros can easily disable the UI or break user-level package installs.
If the distributors will do that, let's do them. Users will decide what they want. We will see.
Sigh I'm right so I'll use my users to force the issue.
Huh? If your users want this disabled or removed - go ahead. Otherwise they may simply not use it. If it really are the users who care about such a feature not being used - who forces them. They may diligently wait for someone to create a package for them. But then that package may use the feature under the hood....
This is all waking up painful memories.
You're right about some users following you. You're wrong about it
having any impact because you won't distribute a proper Linux package,
no self-respecting distribution will follow you
Just like no self-respecting distro allows any software to be installed and used by means other than the package manager....
so they won't too, and the users that try to get things in their own hands usually understand overnight why your proposal is stupid in the Linux context (nothing like doing some package maintenance to understand why others chose these rules)
There are plenty of users who never do any package maintenance, never did and never will. And if an update you do breaks their work environment, they try to fix it by reinstalling the add-on - or complain, if you broke essential facilities for them. Denying them the essential facilities fixes it for you, but not for them.
Before you go this way (and waste everyone's time) don't you think
distributions (esp. not-for-pay) distributions have a little better idea
than you about what their users expect and how to achieve it ?
Sure. Different distros cater for different users having different needs. Because of that they may have to do some adjustments. That includes disabling features their target audience doesn't need. But to me this does not imply that we should decline everyone (and every distro) every feature which some distro will disable.
You know there is no fat contract tying them - they compete solely on their software packaging quality.
Or possibly the feature set of the software they offer. For some customers maximum lockdown is a feature, for others maximum configurability is.
Someone asked the other day why the Ximian fork even existed (and was widely used). This is one big reason. People somehow refuse to accept distribution advice when it goes against common windows wisdom, so distributions have to do their own thing (even Ximian competitors use the Ximian branch)
... but with different patch sets. Not all distros agree on everything. And what works for a Linux distro must not be also correct for Windows or for a standalone, unbundled download.
BTW: A lot of the reasons which brought the Ximian fork into being are already solved. But some - particular those catering specifically for Linux distributors' needs may remain (or take longer) to resolve, because OOo is not just a Linux project.
Ciao, Joerg
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
