1 |
On Mon, 26 Sep 2011 23:06:36 +0200 |
2 |
Jonas de Buhr <jonas.de.buhr@×××.net> wrote: |
3 |
|
4 |
> >The only problem with that attitude is that it eventually leads you |
5 |
> >to the same position that Microsoft is in with Windows -- where too |
6 |
> >many years of refusing to drop backwards compatibility were |
7 |
> >completely holding them back. |
8 |
> |
9 |
> i thought of that too. as with many other things, the trick is to find |
10 |
> the right balance. important code changes/cleanups sometimes have to |
11 |
> be made, even if they break things. if that happens too often its |
12 |
> going to annoy the users. |
13 |
|
14 |
Apple had a nice middle ground, most noticeable in MacOSX. |
15 |
|
16 |
Support the old version in a VM-like environment for two releases then |
17 |
drop the support. I think it's a nice compromise. |
18 |
|
19 |
It's unrealistic to support everything you ever did forever |
20 |
like MS tried to do (IE6 is *still* hanging around somehow...), while |
21 |
the other extreme is probably even worse. The current classic extant |
22 |
example is Amarok2 and kmail2 - in both cases the devs seem to have |
23 |
just decided that anyone running anything older than 6 months isn't |
24 |
worth the effort. Well, that's too bad for Amarok and kmail, there's |
25 |
lots of alternative apps for both. And switching apps is far less pain |
26 |
than trying to deal with upgrades with zero supported upgrade paths. |
27 |
|
28 |
These are hard lessons to learn. |
29 |
|
30 |
-- |
31 |
Alan McKinnnon |
32 |
alan.mckinnon@×××××.com |