Doug posted on Thu, 16 Aug 2012 12:10:35 -0400 as excerpted:
> I have upgraded my "sandbox" from 4.6.5 on PCLOS to 4.8.3, and with the
> exception of a more cluttered systray or "panel" and a somewhat uglier
> desktop, both of which I was fortunate to have been able to revert, I
> see no obvious advantage. (I don't consider the changes I noted to be
> an advantage.) Please explain why I should upgrade my two "work"
> computers to 4.8.3.
> Thanx--doug
An interesting question it is that you ask. From many angles such a
question can be approached.
[How's my attempt at "Yoda-personation"? =:^)
First, as regulars here will no doubt verify, I've made it no secret that
IMO, kde4 only reached what /should/ have been the 4.0 release, with what
was /actually/ numbered 4.5, or more precisely, the latter bug-fix
releases in that feature-release series, 4.5.4 and 4.5.5. Until that
point, I contend that despite kde public propaganda to the contrary
beginning with 4.2, kde4 was simply too feature-incomplete, bug-ridden
and broken to be considered ready to replace 3.5.
While I don't particularly wish to drag up the unpleasant details of that
period, I'm repeating that point here in ordered to make another: Pre-4.5,
there were huge holes in features and functionality, and where things
/did/ work, they didn't always work /well/. As such, until 4.5, the
improvements of each release were huge and dramatic, since in general
whole bits of previously entirely missing or broken but rather basic for
normal use functionality was being filled in.
With 4.5, however, most/all of the basics were in general there and
functional, and improvements since then have been much more incremental
and easier to miss at first glance. In fact, as also well known to
regulars, I'd argue that with kmail's akonadification in kdepim 4.6
(which many distros didn't ship at all, sticking to updates of kdepim 4.4
until some time in 4.7), kdepim actually reverted to a broken what should
have been pre-release state once again, causing many users including me
to switch to something else for mail. Personally, I've been extremely
impressed with and recommend claws-mail, tho those who prefer HTML mail
and/or pim-suites that include calender/scheduling, feed-reader, etc, are
unlikely to find it suitable.
>From that angle, pretty much any kde4 since late 4.5 has the necessary
basic functionality, tho it's worth noting that 4.5 itself was still
hal-based for hardware management, while 4.6 switched to udev/udisks/
upower, and most distros dropped hal at about that time as well, so 4.6
is really the minimum for a post-hal system (unless that functionality is
disabled at compile time, a choice gentoo, for instance, as a scripted
build-from-sources distro and my distro of choice, provides).
So from that viewpoint, anything from late late 4.6 on (early 4.6 had
rather severe issues, in part due to issues in kde's general switch from
svn to git at about that point), still with the kdepim 4.4 if you install
any kdepim at all, should suffice.
But as I said, incremental improvements have continued. You might not
notice these immediately, in much the same way as you don't necessarily
notice it the moment a headache goes away, and only a bit later realize,
"hey, I don't hurt any more!" In fact, if you don't customize much, you
may not notice many of them at all (but in that case, honestly, a lot of
folks prefer something "simpler", with less "confusing" options,
something that "just works" as-is, without too much chance to change it
and perhaps break it, or at least that's my opinion... as a generally
heavy customizer who never met a desktop he liked as-shipped and doesn't
expect he ever will!). A lot of it there depends on what kde components
you actually use, and as I said, often on how much you take advantage of
kde's ability to customize.
Two cases in point have been quite noticeable /incremental/
/improvements/ for me, even if 4.5/4.6 /was/ generally usable: (1) kwin's
window rules, made much easier to manage in 4.7, and (2) konsole
functionality, with a number of very long-term bugs fixed in 4.9 (which
with 4.8 you won't see yet), some of which went back to the kde3 era,
before kde4 at all!
Kwin's window rules dialog is where you go if you have a window that you
want to behave differently than it does -- for instance, perhaps you want
it to consistently start at the same size, even if you resized it the
last time you used it and it normally remembers that, so would otherwise
start at the newly resized size it remembers instead of the consistent
size you want it to start with. Or maybe you want a particular window to
always appear maximized... Whatever the case, the window rules options
can be reached in two different ways, either from the window menu
available from any window (context-click on the window decoration, or
there's a hotkey, alt-F3 IIRC by default), or from kde settings,
workspace appearance and behavior, window behavior, window rules (that's
the 4.9 location, I think it might have changed slightly, recently).
I have... eyeballing... about 30 of these rules setup for various things
so I use window rules enough to have noticed this within a couple of days
after it was introduced in the 4.7-pre-releases I had decided to test.
The window specific settings dialog you get if you setup a rule is quite
a bit easier to understand and work with in 4.7+ as compared to 4.6 and
earlier. Previously, there were two different tabs that affected rule
matching, in addition to the tabs that setup what the rule actually did.
With 4.7, the two matching tabs were combined into one, and the wording
of both the matching options and the various actions was changed to be
MUCH clearer and easier to understand. With 4.6, it was much more
difficult to actually setup a rule to do what you wanted, because many of
the options were unclear. In 4.7+, the wording is far clearer, and in at
least two cases (ignore requested geometry and obey geometry
restrictions, there's now a tooltip with an additional explanation, if
you hover over the option. I know /I/ never understood the difference
until I had those tooltips to read!
That's exactly the sort of "incremental improvement" that kde has been
getting a lot of, since 4.5/4.6. Not visible right away, and not
something everybody's going to notice at all, but it really does improve
usability. =:^)
Similarly with the konsole improvements and bugfixes in 4.9. There were
quite a number of "little" bugs fixed, including as I said, some opened
against konsole for kde 3.5 or earlier. The one that most affected me
had to do with how the mouse works in the konsole terminal window.
Original Linux (Unix?) terminal mouse protocol limitations going back to
the age of 640x480 screens at 80 columns by 25 lines didn't allow for
proper mouse output on today's widescreen monitors, often 1920 pixels
wide -- clicking in a spot that happened to be too far right on a
widescreen would produce gibberish output. (I think the same applied to
too far down, as well, but am not sure.) There's two protocol extensions
available that can fix this problem, with fixes applied to some of the
other terminal programs, but not to konsole until 4.9.
Again, that's just one of a whole list of bugfixes that hit konsole with
4.9, but it's one that made a big difference for me, since I regularly
use mc (midnight commander) as a "semi-gui" file manager in both text VTs
and in konsole under X/kde. Again, this is /exactly/ the sort of
/incremental/ improvements that various kde components have been getting
since 4.5/4.6, improvements that you may not even notice, but that
together, make for a much improved kde experience, in general.
So that's another angle...
A third angle is security related. If you're the only user of your
computer, and don't use kde tools (konqueror, kmail, knode, kopete,
akregator, etc...) on the Internet so don't have to worry so much about
security vuln exploitation from the net, you /might/ find yourself able
to ignore security issues, tho I could never recommend it unless you
REALLY know what you're doing, in which case you'd likely not be asking
questions such as the above.
Actually, while we're on the topic, I'd recommend NOT using konqueror for
a primary web browser in any case, as KDE devs themselves obviously
consider it only a toy. What else would you call ignoring the fact that
in kde4 konqueror was left without security cert management, at a time
when whole certificate authorities were having their certs revoked, for
YEARS after kde4 was claimed to be ready for "ordinary users" (with 4.2),
when in fact, such "ordinary users" could easily be depending on their
browser for web-based banking, shopping online, and the like? (I know I
do nearly all of my banking online here, tho I guess I'm probably a bit
beyond "ordinary".) Additionally, one of those early 4.6 issues I
mentioned above was a "double form submission" bug in konqueror -- you
know those financial transaction sites that specifically warn not to hit
the submit button twice, or you might be billed twice?... While such a
bug would have probably never made it out of beta in a /decent/ browser,
it was actually introduced in kde 4.6.2 IIRC, a bugfix release that as
all bugfix series releases, was claimed to be safe to upgrade to, as it
supposedly contained bugfixes only. But, not only was it introduced in a
bugfix release, but kde waited TWO MONTHS before releasing the fix in
4.6.4, 4.6.3 didn't fix it. Compare that to any of the big browsers,
which would have almost certainly had a high priority fix, especially as
the commit introducing the problem was known, out in days!
So either kde devs are entirely irresponsible, and deliberately recommend
products such as konqueror that they know don't get proper security
support for use by "normal users" for "normal usage" such as online
banking, etc, or... they simply don't consider konqueror any more than a
trivial toy... not used for "important" stuff like ordinary online
financial transactions, by ordinary users.
Either way, the effect is the same, konqueror is a definite WARN on my
list in terms of more than trivial online use. Strict security practices
would insist it shouldn't even be installed, but I /do/ still keep it
installed here. I just set kde's browser pref to firefox, and rarely use
konqueror at all any more. I'll probably uninstall it at some point, but
haven't yet.
Then there's all things kdepim, including kmail, akregator, knode, etc.
The kdepim devs are integrating all kdepim apps with akonadi,
"akonadifying" one app at a time, starting with kaddressbook in IIRC
4.4. I already mentioned that with kmail akonadification in kdepim 4.6,
it regressed to pre-release quality.
To be fair, since I migrated to claws-mail before installing 4.7, I don't
know from personal experience how kmail has done since. But reports say
that at least as of 4.8, it was still eating people's mail occasionally,
and causing less severe issues more frequently. So while they may
eventually fix it, if you use kmail, I'd definitely recommend NOT
upgrading it past kdepim 4.4.x (kmail 1.x, 2.x is the akonadified version
that started shipping with kdepim 4.6), tho as I mentioned, many distros
continued to ship kdepim 4.4.x thru kde 4.7. And I'd suggest switching
to something else when you otherwise would upgrade kdepim beyond 4.4
(which again, depends on distro).
What I actually did here was uninstall all of kdepim entirely, tho the
only apps I had been using from it were kmail (and kaddressbook of
course) and akregator, so that's all I had to find replacements for. But
then I could uninstall it, and along with it, akonadi.
So for anything kdepim including kmail, I'd actually recommend switching
AWAY from kde/kmail/kdepim, to something else. As I said, I found claws-
mail a good match for me here and can recommend it for those that don't
need a pim-suite and who prefer plain text to HTML mail. Otherwise,
consider thunderbird or evolution or something, but I'd still say stay as
far away from kmail2/kdepim-4.6+ as you can possibly get! Paraphrasing a
well known saying: "The mail messages you save might be your own!"
But other than anything kdepim (and other than konqueor as a browser),
I've actually been quite happy with newer kdes, and really have found the
incremental improvments to be just that, /improvements/. =:^)
As for the two specifics you mention, systray and desktop, those are
/entirely/ configurable, and as long as you save /home during upgrades,
once you've customized them, they should stay pretty much the way you put
them. It's only if you leave them at the defaults that they're likely to
change, since the defaults do change sometimes. They usually call the
default changes an improvement, but as I said I've never seen an as-
shipped desktop that I was happy with and don't expect to, either, so I
don't worry too much about /what/ the defaults are, as long as they don't
mess with my customizations, and I'm happy to say that for the most part,
barring the occasional bug, they've been pretty good about leaving those
as they are. =:^)
Bottom line: From late 4.6 the improvements have been mostly
incremental. If you don't see a reason to upgrade, and you are in a
position where you don't have to worry about security vulns, then really,
unlike some commercial "subscription software" plan where you have to
stay with what they're supporting or it ceases to work at all, it
shouldn't actually /break/, and you should be able to continue running
the old versions as long as you want, at least until your hardware breaks
and you find the new hardware won't work with the old distro version you
have.
But regardless, I'd definitely recommend switching to something other
than konqueror for browsing, as that's a potential security issue, and
I'd definitely recommend either sticking with kmail1/kdepim-4.4, or
switching to something else, as the messages you save may indeed be your
own!
--
Duncan - List replies preferred. No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master." Richard Stallman
___________________________________________________
This message is from the kde mailing list.
Account management: https://mail.kde.org/mailman/listinfo/kde.
Archives: http://lists.kde.org/.
More info: http://www.kde.org/faq.html.
