<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<p><br>
</p>
<div class="moz-cite-prefix">On 22/10/2019 23:22, Sean Davis wrote:<br>
</div>
<blockquote type="cite"
cite="mid:2ba2203e-188a-4e9a-b77c-4c12893154b4@www.fastmail.com">
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
<title></title>
<style type="text/css">
p.MsoNormal,p.MsoNoSpacing{margin:0}</style>
<blockquote type="cite" id="qt">
<p>If you really want work on that, fine, but as a minimum
please make it a run-time option and make sure the non-CSD
variant always works. Otherwise, if you really must follow
through, rename/fork the applications and libraries so that
original can be maintained and developed further under their
current names and in their current repositories. We don't want
to end up with another Gnome/Mate mess.<br>
</p>
</blockquote>
<div><br>
</div>
<div>Another feature of the toolkit CSD decorations implementation
is that GTK_CSD=0 just works. The last thing we want is any
divide where Xfce is forked and maintained separately.<br>
</div>
</blockquote>
<p>This needs to be an option in GUI and CSD should be off by
default. Otherwise, *please*, fork the code. One of the problems
with Gnome3 was they have have repurposed Gnome naming and
infrastructure to push their ideas and people who wanted to use
the previous implementation were forced to do the fork.<br>
</p>
<blockquote type="cite"
cite="mid:2ba2203e-188a-4e9a-b77c-4c12893154b4@www.fastmail.com">
<blockquote type="cite" id="qt">
<p>But, why? Why break established X11 patterns for the sake of
some visual candy? Window managers are a central part of X11
user experience. Only recently my organization has deployed
Xfce on user machines because the hits a good balance between
being standards compliant and feature rich. From the usability
point of view CSD have a negative value.<br>
</p>
</blockquote>
<div><br>
</div>
<div>It's not about the visual eye candy. There are actual
benefits we gain from CSD (as I mentioned before), and the
window manager is still at work when using the CSD windows.</div>
</blockquote>
<p>That's all great. Now, make it an option or fork the code -
that's the way to convince users.<br>
</p>
<p>To me, this feature has a overall negative value and there were
many technical problems with it listed as well. Two main issues to
me:</p>
<p>- We lose credibility as a project. The community *will* see it
as Xfce team turning into a Gnome3 path. This is particularly
damaging to us because many of our users came here in search for
an alternative to Gnome3.</p>
<p>- It is simply not possible to implement all features of all
window managers, so this will always lead to loss of functionality
and compatibility issues. By all means, do try to minimize them,
but also don't force users to use it in the first place.<br>
</p>
<p>By the way, if in future you want to work on a Wayland compositor
- that's OK, just make it a new tool, don't just replace
xfwm/xfdesktop/xfce4-panel. In that case it is probably OK to
default to CSD, or at least have it supported by libraries,
although an option for having SSD would be great.<br>
</p>
<p>Best regards,</p>
<p>Andrzej<br>
</p>
</body>
</html>