In comp.os.linux.advocacy, unionpenny@xxxxxxxxx
<unionpenny@xxxxxxxxx>
wrote
on Tue, 30 Oct 2007 12:15:21 -0700
<1193771721.637558.273690@xxxxxxxxxxxxxxxxxxxxxxxxxxx>:
> On Oct 30, 11:43 am, Erik Funkenbusch <e...@xxxxxxxxxxxxxxxxxxxxxx>
> wrote:
>> On Tue, 30 Oct 2007 12:09:11 +0100, Jim Relsh wrote:
>> > "Roy Schestowitz" <newsgro...@xxxxxxxxxxxxxxx> schreef in bericht
>> >news:4409392.pQp5N9ZnWq@xxxxxxxxxxxxxxxxxx
>> >> Miguel's delusions of grandeur
>>
>> >> ,----[ Quote ]
>> >>| What is Miguel de Icaza's latest game? His obsession
>> >>| with tailgating everything that Microsoft develops -
>> >>| and trying to impress the company by producing Linux
>> >>| equivalents - now seems to be spreading to others as
>> >>| well.
>> >> `----
>>
>> > I agree with the author. Miguel is too smart to do such stupd
>> > things without some ulterior motive.
>>
>> .net is a *good* technical solution. Miguel sees that.
>> Everyone is jumping down his throat for non-technical
>> reasons. Not a single person has come up with any valid
>> criticisms of .net from a technical perspective.
>
> But is it good and compatible for a GPL project?
Define "good and compatible".
The main technical issue I have with .NET is a
performance-related one; it is far from clear which of
the static optimization offered by .NET's compiler/backend
and the dynamic optimization Java's runtime supports works
more effectively. Counterbalancing this is .NET's richer
service support -- SOAP is built into the library, for
example, and is easily set up. Both are multiplatform,
and .NET has more languages.
There are a few non-technical issues, of course, most of them
having to do with patenting, arbitrary Microsoft revisions,
and general compatibility with non-Microsoft solutions.
Also:
- can a Microsoft-built application run on a non-Microsoft OS
on Microsoft-compatible hardware? If so, how?
- can a Microsoft-built application run on a non-Microsoft OS
on non-Microsoft-compatible hardware? If so, how?
- can a non-Microsoft-built application run on a Microsoft OS
on Microsoft-compatible hardware? If so, how?
I'll admit I prefer Java, but Java has many issues and a
few licensing quirks. The issues are going to be hard
to address without major modifications, either:
- int vs Integer et al (the primitive versus object issue;
boxing/unboxing is a workaround but hardly a fix)
- .size() versus .length versus .length()
- removal of List.get(int); granted, it's useful
- (new Object[n]).iterator() as opposed to
Arrays.asList(new Object[n]).iterator()
- foreach(Object o : (new Object[n])) { ... }
- metadata retrieval from a Class descriptor (this might
already be doable but I'd have to look)
- proper template instantiation and verification (the
current implementation is a bit of a hack)
- problematic locale handling; e.g. reading news in Java
could lead to some interesting issues if some post
using iso-8859-1 and others using utf-8 -- best I can
do here is read the socket a byte at a time looking
for newlines and the '\n.\n' or '\r\n.\r\n' sequence,
and that's yugly with a capital ugh
- timezone issues (EST being the worst offender)
- replacement of the proprietary binary RMI protocol with
SOAP/XMLP
- better source code formatting from javadoc
etc.
These aren't fatal but they are annoying.
Not sure what quirks C# has apart from the interesting issue
of what happens in an expression such as a[b] = c[d] when
a and c are classes with property setter and getter methods.
>
>
>> Just goes to show, make the wrong political move in the
>> Linux camp and you get lynch mobs.
>
--
#191, ewill3@xxxxxxxxxxxxx
Windows Vista. Because a BSOD is just so 20th century; why not
try our new color changing variant?
--
Posted via a free Usenet account from http://www.teranews.com
|
|
