Re: soapsuds for binary

Tech-Archive recommends: Repair Windows Errors & Optimize Windows Performance

From: John Wood (spam_at_isannoying.com)
Date: 06/21/04 

Date: Mon, 21 Jun 2004 08:34:03 -0400

Thanks Sahil. I always find Ingo's perspective interesting (I have most of
his books), although I don't always agree with what he says. For example,
the whole concept of passing a session ID to each call in an SAO singlecall
seems very low level and is something that should be handled by the remoting
framework. In all the time people have chatted about this, someone could
have just written a custom channel that does all this transparently (making
singlecalls behind the scenes with session passed in transparently).

He talks about using interfaces to access remote objects -- which makes
sense, but why interfaces and not abstract base classes? Interfaces are
meant to abstract access to a particular aspect of an object's
functionality. When you're talking about abstracting the interface to an
entire object, it seems more appropriate to be using abstract base classes.

Now I agree with the concept he is discussing, but still disagree with his
premise that it wouldn't be possible to auto-generate these interfaces /
base classes. I agree there are problems with chatty functions, but a good
code-generator should provide you with a template - a starting point - and
should allow you to fiddle with the output without any risk of losing those
changes in the next generation.

John

"Sahil Malik" <nospam@ms.com> wrote in message
news:%235Ziid0VEHA.1652@TK2MSFTNGP09.phx.gbl...
> It cannot be .. here is a good discussion on that in Ingo's weblog (in my
> eyes the remoting guru)
>
> http://www.thinktecture.com/staff/ingo/weblog/archives/001249.html
>
> --
> - Sahil Malik
> Independent Consultant
> You can reach me thru my blog at -
> http://www.dotnetjunkies.com/weblog/sahilmalik/
>
>
>
> "John Wood" <spam@isannoying.com> wrote in message
> news:ut$HwNVVEHA.3988@tk2msftngp13.phx.gbl...
> > Well I was kinda hoping that the auto-generation would generate both the
> > abstract base classes and the derived implementation classes, but only
if
> > the derived implementation class isn't already defined. This provides
you
> > with the opportunity to customize the proxy class yet continue to reap
the
> > benefits of auto generation.
> >
> > I really want to avoid hand coding proxies, because it is a lot of
manual
> > labor that in this day and age really should be automated IMHO.
> >
> > "Allen Anderson" <allen@sparkysystems.com> wrote in message
> > news:u866d09g18vqs85ejonh663cebju97pq7t@4ax.com...
> > > sure, let me be a bit more clear on that. Auto generation is not in
> > > and of itself a bad thing, especially when coming to DB operations (I
> > > use them too). I'm only speaking in the context of remoting using
> > > soap suds. I used to use soap suds myself for all my class
> > > generation. I found that because things were being auto generated
> > > instead of being controlled by me, that I lost the ability to
> > > customize the proxy. Once you begin writing your proxies as interface
> > > or abstract class you become more aware of opportunities to write
> > > things in at the proxy level where they didn't need to be fully
> > > remoted. It's one of those things that after you have done it a lot,
> > > you start seeing more and more.
> > >
> > > Defining your proxy also allows you to hide methods on your server
> > > class that you may not want to expose otherwise. The devil is in the
> > > details I guess and the more you use user defined proxies the more you
> > > see their necessity.
> > >
> > > Cheers,
> > > Allen Anderson
> > > http://www.glacialcomponents.com
> > > mailto: allen@put my website url here.com
> > >
> > > On Fri, 18 Jun 2004 12:24:00 -0400, "John Wood" <spam@isannoying.com>
> > > wrote:
> > >
> > > >>> Auto genertion leads to all sorts of potential problems and
general
> > > >sloppiness.<<
> > > >I'd be interested in hearing more specifics of that. I've been using
> auto
> > > >code generation for data-tier access for years and have found it to
be
> > > >extremely useful and time saving.
> > > >
> > > >"Allen Anderson" <allen@sparkysystems.com> wrote in message
> > > >news:dk46d05n17j8jvg5e0cjtt1q8bq38uf3mf@4ax.com...
> > > >> not that I'm aware of. However, just as a note, I wouldn't use
> > > >> automatic programs to generate proxies anyway. Your much better
off
> > > >> defining an interface or abstract class to serve as the proxy for
> your
> > > >> object. Auto genertion leads to all sorts of potential problems
and
> > > >> general sloppiness.
> > > >>
> > > >> Allen Anderson
> > > >> http://www.glacialcomponents.com
> > > >> mailto: allen@put my website url here.com
> > > >>
> > > >>
> > > >> On Fri, 18 Jun 2004 11:05:12 -0400, "John Wood"
<spam@isannoying.com>
> > > >> wrote:
> > > >>
> > > >> >Is there anything like soapsuds that generates proxies for binary
> > > >remoting
> > > >> >rather than soap?
> > > >> >
> > > >>
> > > >
> > >
> >
> >
>
>

Relevant Pages

  • Re: separating specification from implementation
    ... interfaces or abstract types because method calls are resolved by name ... advantage over abstract base classes that there is no requirement to ... Ruby is a dynamic language whereby a method call is resolved at runtime ...
    (comp.lang.ruby)
  • Re: soapsuds for binary
    ... As a support for Allen's answer, you may take alook at this Ingo ... but why interfaces and not abstract base classes? ... it seems more appropriate to be using abstract base classes. ...
    (microsoft.public.dotnet.framework.remoting)
  • Re: soapsuds for binary
    ... but why interfaces and not abstract base classes? ... >functionality. ... it seems more appropriate to be using abstract base classes. ...
    (microsoft.public.dotnet.framework.remoting)
  • Re: Comparing abstract base class and Interface
    ... You have probably heard the saying "Program to interfaces, ... However, you often meet abstract base classes in frameworks where some default behavior is needed. ... interfaces are Java's way of providing a multiple inheritance feature. ...
    (comp.lang.java)