Everhart, Glenn
From:	Gregory Krohne [GregoryKrohne@worldnet.att.net]
Sent:	Monday, August 10, 1998 4:58 PM
To:	Info-VAX@Mvb.Saic.Com
Subject:	Re: Porting/integrating OpenVMS & Windows NT

JOHN MALMBERG wrote in message <0095000016451823000002L032*@MHS>...
>>As I understand it, the plan here is to alter the application just enough
to
>>compile and run under NT and OpenVMS. This way, ongoing features and
patches


Just so.

>The WIND/U stuff from Bristol Technology allows you to do that.  License
fee
>can be expensive.

I looked them up. Here's the original announcement:
http://www.bristol.com/Company/Announce/digital-dcom.html
I want to go from OpenVMS to NT, but Bristol at least offers a platform
where I can run the same code on both machines. Since Wind/U does cover
thread processing and memory-mapped files, it's worth a look.

I did check out Sector7:
http://www.sector7.com
They say all the right things regarding porting code from OpenVMS to NT (and
Unix, as well). And while the tools for porting to UNIX have been available
for some time, the tools for porting to NT are apparently not finished yet.
As in, they're almost ready for Beta. Sector7 does offer services where they
will port the code for you, but their fees for this service are high. Of
course, you and your company may find Sector7's services a bargain.


>It would actually mean porting entirely to NT, and then backporting to VMS.


Yes, as long as we're talking about such big changes, why not just rewrite
it all in Java? I know it sounds strange, but at least Java code supports
multi-threaded processing right in the language. So, we could write
multi-threaded applications that should run equally well on NT and OpenVMS,
without code changes between the two. Any thoughts?

>There are also some competitors to them, but I do not have a list handy
here.


Do post a list, if you find competing products.

>Also check the Digital WWW site.  They have some enhancement package for
the
>DEC C compiler for WIN32 compatibility.
>
>In the cross porting I have done, I have found that the VMS compilers,
Linker,
>and operating system trap and accurately report where many source errors
are
>that the other platforms just indicate by a random crash or data
corruption.
>With WINDOWS programming you may need an ICE unit to get that detail.

>
>So building a VMS version still has quality advantages to you.

Just so.

>The next thing is, when they go to NT, are they going to be still using
Compaq
>hardware?


I don't know. Probably, just to get the trade-in value on the old VAXen,
anyway.

>Will they be using ALPHA?

No decision on that, yet. Again, I think the trade-in value of the VAXen
will figure into it.