On 9 Mar 2006 at 16:11, Pascal Bleser wrote: [...]
And although I strongly dislike Mono and even more .NET, it does make sense.
I mean, it's also to consolidate efforts. Why should SLES and NLD (I presume) use Zenworks and SUSE Linux not ? Concentrating the efforts and the manpower on one engine is beneficial to everyone, both to Novell's budget and to end-users because you don't split and duplicate efforts.
The question is: How many extra MB does this bloated framework cost on the harddisk? Is it definitely free from patent claims from Microsoft?
Now the fact that it is based on Mono.. well... I don't care much. It's an application, that's it. As long as it works and people take care of proper integration and functioning, I don't care.
The more code, the more bugs it seems.
The only problem I have with Mono is more or less the same as with GNOME in general: it is not stable as a platform, not by any means. Almost every single piece of OSS project out there that uses Mono cannot run on SUSE Linux 10.0 (that, arguably, is not very old) because it requires the very very latest version of Mono. Note that the same very often applies to GNOME (much less with 10.0 as it ships a very recent GNOME). That is proof to me that either the engineering is quite poor or (more likely) that Mono is still far away from being a stable platform, feature and API wise.
That just another dependency. For HP-UX you end up with having several versions of JRE installed. Some applications (like Oracle) ship their own JRE, just as some applications do. I hoipe will not end up in solutions like that.
But it also shows that Novell still has some integration work to do on its products. While the former Ximian folks keep on coding on Mono and GNOME (and the gazillions of libraries and subprojects that relate to it), that stuff is not available on Novell's own distribution. I find that rather pathetic (why would Mono or GNOME work better on Ubuntu or Fedora than on SUSE Linux ?) but well, let's hope it gets better, and I presume it will with the release of 10.1 that includes the very latest stuff of almost everything.
Everytime I thought there is a suable library for the functionality I needed, I realized that it is either - not documented in a way that it could be used - very restricted in it usefulness so I ended up writing my own implementations. And for GNOME as seen in SuSE Linux, it seems just to be cool to fill the logs with failed assertions. Failed assertions are fatal software errors that I never want to see in a released product. Maybe the mentality is like this: We know that the code is not quite correct, so let's put an assertion here to get reminders of the problem. I'd perfer very much fixing the problems instead of add ing assertions that are meant to fail.
There are for sure integration issues, which we will hopefully address in time. The SL 10.1 distribution will use rpm-md+ repositories I have just heard, so smart (?) yum (?) will work fine too for updating 10.1. (This is not finalized yet.)
Yes, that's a very good feature. While yast2's own repository metadata format has proven itself to be reliable and work well for years, it is not particularely well documented (AFAIK) and exclusive to SUSE Linux (and SLES and NLD), which is definitely not a good thing. Fortunately Mauricio Teixeira has implemented yast2 repository support in smart so we do have another tool than YaST2 as a package management frontend. That is good because although YaST2's package management module is stable and quite easy to use, it sucks at a lot of things, e.g. solving dependencies and conflicts (smart is very good at that), using mirrors or managing 3rd party repositories (wtf are those locked packages ?).
Hopefully the switch to the Zenworks engine will bare improvements in that area.
Is there some guarantee that Zenworks is free from patent claims (from Novell)?
It doesn't really matter if Novell uses the MS .net "initiative" for own
It's not the ".NET initiative", it's Mono. Forget about Mono being .NET because it isn't. Miguel can keep on trying to make everyone believe it us, it's more or less bullshit IMVHO. It's just Mono. Whether that is good or bad is a topic on its own, personally I don't care. It's a bytecode-compiled, fast and modern language (although it has a few very questionable architectural choices but well, what else would you expect from microsoft). As long as it works... And honestly, I can't be a lot worse than C or C++ for that kind of application.
Fortran programs did work quite some time as well. "it works" is not a quality indicator for me. "Assembler also works".
software products in a mixed environment. However, if software updates and distribution for any kind of linux product will _depend_ on mono or any other linux implementation of .net, people like me (former MD of Gartner Group) will help kill Novell Linux in marketplace. And very efficiently, believe me.
Wow, you've just totally disqualified yourself, both by threatening and even more by stating you've been working for that bunch of completely technically unsavvy people who are driving the IT market nuts by making utterly stupid and unfunded statements that most IT managers run after (and pay a lot of $$$ for). Mate, you're part of the hype curve.
Ulrich