[ILUG] Re: [ILUG-Social] GNOME to "embrace & extend" .NET !

[Padraig Brady]

I think everyone should adopt .NET and not go off designing
something different. To me .NET is required as the (RPC)
transports used by COM/CORBA required loads of open
ports and also connections in both directions which are impossible
really to adequately firewall. All the info is now nicely spit over an
XML stream which does require more CPU processing than the
binary RPC protocols but everyone has loads of that. I don't
see any disadvantage for easing the interoperability between
logic on various systems. Also it fits more easily with the Unix
philosophy of cohesive lumps of logic connected with text
streams, in my opinion.

As for the logic connected by these XML streams, it doesn't
matter that it's implemented in C/C#/Java/Python/... If you don't
need to distribute platform binaries as for C#/Java/Python/...
it makes things a little easier for all involved, but it's not major
in my opinion as it's easier to connect to different platforms
anyway, and also binary distribution systems are getting much
better lately.

As a side note I think the various M$ database APIs were ridiculous,
DAO, RDO, ADO1, ADO2, ADO+, .. they keep changing the
interfaces every year it seems like.

Padraig.

Glen Gray wrote:

>Lets face it, we all new this was coming and it really does have to
>happen. The whole idea of Bonobo etc. was all leading to this. And for
>anyone who got to speak with Micheal Meeks at last years Linux World
>after party here at antefacto or hear his speech at the conference can
>tell the kind of attitude Ximian have to .NET. If your not in you can't
>win. And they see .NET as a winner and a killer if there isn't an
>alternative.
>
>Having spent most of the last 5 years working on Client/Server
>applications under the Microsoft technology umbrella and seeing how
>popular that approach is with the majority of mid-sized business (which
>is probably the largest sector) I'm amazed that there isn't the same
>level of "ease of development" available for linux. Now before you all
>go mad and flame the shit out of me hear me out. 
>Microsoft has spent a lot of time and effort pandering to the needs of
>developers and doing lots of hand holding. Early access to new
>technology and tonnes of samples and write ups via MSDN (which is a
>wonderful service). Stuff like ODBC came along to make life easier when
>accessing databases. They then took that further with a few different
>technologies and have settled with ADO+. I don't want to get too caught
>up in all of this but the point I'm making is that there are a LOT of
>developers out there that depend on the levels of abstraction that
>Microsofts technologies provide to enable them to build applications
>quickly and easily. Ximian also see this and are trying to brin GNOME
>into that field of play. It was one of the design goals of GNOME to
>allow this to happen.
>
>Now whether your personal opinion is that these people aren't really
>developers if they need to work in environments like that or that
>Microsoft are wrong to do those things is irrelavant given the fact that
>there is such a large portion of the development community out there
>that are just as I described. I personally think they are doing good
>things in certain areas but frustratingly they can still make a balls of
>it e.g they've gone through several versions of the COM architecture
>over the last 8 years or so, mostly each release required massive
>changes to code to keep up with the technology and often meant adopting
>new versions of other technologies, OLE became COM which blened heavily
>with RPC to become DCOM which progressed into ActiveX which evolved into
>COM+. ODBC abstracted to RDO which branched off to ADO and now has
>evolved to ADO+ and I think that RDO died at ADOv1.2. They tend to
>promote technology for adoption that just isn't ready and just as it
>becomes stable and usable there is something else being pushed to
>supersede it. This is really annoying when you are working on really
>large projects with long timescales, when you've finished what was a
>cutting edge project it's not quite as shiney as when it was designed 9
>months ago. It makes it hard to re-use code too as there are always
>subtle changes in the way the COM interaction is done or DLL interfaces
>change a little. 
>
>I'm seeing .NET as a maturation of alot of what Microsoft has been doing
>with these technologies over the last few years. Their converging into a
>single "mission statement" type of scenario where there is actually a
>unifying purpose to all this COM+ and ADO+ technology. Your not only
>using what Microsoft are claiming to be the new wave of tech but they
>are too, in a much more visable manor. Now these technologies aren't
>neccesarily anything new or great. CORBA already did most of what
>Microsoft wanted when it came to designing COM, but true to their
>strategies, they took from it what they wanted and went and did it their
>own way. Lets face it though, everyone does that kind of stuff, Bonobo
>is heavily based on where COM+ is now. But Microsoft had a direct
>channel to developers via MSDN and pushed the adoption of these by
>developers and that approach has served them well.
>
>What I'm mainly interested in is what Microsoft is going to do about
>projects like Mono. On one hand they've seen the flaming they get from
>the general IT public and media for their Embrace and Extend tactics and
>are trying to get this framework accepted by pushing it as an open
>standard. But how are they going to turn it to a propietry framework /
>defend their market where they'll make their money as a service portal.
>From all the press snippets I've picked up over the last year or so the
>whole .NET machine that MicroSoft has put in motion is effectively a
>change in direction of their business model. They are effectively
>turning themselves into THE software services company. We've all read
>the recent shifts in attitude to what the license entitles you to. Newer
>version basically will turn into a rental agreement where you don't own
>your version of the software, you are merely renting a service from
>Microsoft. They seem to be heading for the Sky Box Office approach to
>software. It seems that alot of eggs are being placed in the .NET
>basket.
>
>I think it's safe to say that subtle changes in protocols and the
>Microsofts version of "Embrace and Extend" may resurface at a later
>stage. How successful those tactics will be this time around will be
>questionable. IIRC they didn't get away with it using Kerberos.(What
>happened there, did we all have to adopt their extensions or did they
>revert back to kerberos standards ?). But then they are such a large
>company that they will be able to refocus if/when things don't go their
>way with .NET. 
>
>Sorry that turned into such a long one, I don't write often, but when I
>do I tend to jabber.
>
>Glen
>
>
>