I hesitate to jump in on this but, I must.
Frameworks have been used in virtually all areas of software development to
save time and effort and thus save money. For example MFC and ATL have been
used with great success for Windows user-mode applications. Now .NET is
doing the same for both user-mode and web applications. Obviously
frameworks are seen as valid by the software community as a whole. Driver
frameworks now seem to be supported by Microsoft as they are busy developing
their own.
Looking at some past successes, MFC is a buggy bloated pig. But, compare
writing a sufficiently complex application using the Win32 API alone versus
using MFC, and MFC doesn’t look so bad. Obvioulsy many developers have
opted not to use MFC, but to say that MFC has not saved numerous
corporations numerous man hours and been a net benefit overall, I think is
to mistate the reality of the situation. MFC, while maybe not the best
solution, has allowed the software industry to move forward more quickly
than it otherwise would have. I realize MFC is obsolete today, I am just
using this as an example.
Now, many developers may have issues with using a framework in the kernel.
And on public newsgroups you are much more likely to hear from detractors
than supporters of frameworks. This is simply a matter of economics.
Consultants make money because Windows drivers are sufficiently difficult to
develop. If tools make driver development easier, that isn’t good for
consultants. Further, these boards are heavily populated by consultants as
these boards are an excellent advertising medium. Believe me, I have no
problem with consultants whatsoever, and there are many talented consultants
here. However, these are the facts.
That said, there are also many fly-by-night consultants, and everyone likes
to think they are an expert. Most good consultants, while they likely will
steer you away from frameworks for obvious reasons, will not bash tools that
have been helping many many companies for many years. The sorts of
statements you say your consultant is making smack of a lack of experience
in this industry frankly.
DriverWorks has bugs. I wouldn’t try to deny this for an instant. But
guess what, so will the driver that your consultant writes. However, our
framework code has been run by literally thousands of users on literally
thousands of different hardware platforms. I wonder if your consultant can
claim the same? If you take even a cursory look through the DDK samples,
you will find plenty of bugs there as well. Does that mean the DDK samples
are bad? Maybe some of them. On the whole I wouldn’t say that is true.
Every significantly complex software project is going to have some bugs.
However, I believe most of the long time contributing consultants on these
driver related newsgroups will write you a solid driver no matter the
application. Bug free? Probably not, but as close as you will come. In
the same vein, I believe DriverWorks does and has provided solid quality
driver code for its users. Bug free? No, but I do believe the quality to
be very high. Certainly as high or much higher than someone who has not
been writing Windows driver for several years could hope to produce.
Okay, maybe some of my opinion crept in here so take it for what it is
worth. But, weigh your consultant’s statements as well. He is every bit as
biased as I am. I think either path consultant or framework can produce the
same results given the requirements of the particular application. They are
both tools. Use them properly and you will have good results.
Hope this helps.
–
Bill McKenzie
Compuware Corporation
Watch your IRPs/IRBs/URBs/SRBs/NDIS pkts with our free WDMSniffer tool:
http://frontline.compuware.com/nashua/patches/utility.htm
“Ilya Dreytser” wrote in message news:xxxxx@ntdev…
>
> Hi.
>
> My company is using the DriverWorks package to create several PCI drivers.
>
> Drivers created, drivers work, everything is great!
>
> Then our client gets us to hire a driver expert consultant, who proceeds
> to tell them that DriverWorks is very very bugy, poorly supported, and not
> fit to write a driver for a toaster!
>
> So, I wanted to get the comments of other people in the driver development
> field, the real life if you will (since this guy basically teaches
> Mircosoft driver writing for a living), and see what people say about
> DriverWorks, and their experiences with the package.
>
> Thanks!
>
>
