Forum OpenACS Q&A: OpenACS vs .NET: How is the pitch made?

I will soon be walking into a meeting where the chief technical consultant is leaning heavily towards a .NET front end application stack in front of an Oracle db running on Solaris. They aren't near making the decision yet, but the battle for mind share has begun.

Since the .NET juggernaut has really begun to gather steam (just check the classified ads to see that .NET projects are getting as popular as J2EE ones), I was hoping the community can help me map out some comparisons of the OACS stack vs .NET.

There are of course some obvious ones (BILLIONS OF DOLLARS!!!) but I am more interested in the technical arguments. The issue of monopoly vs OSS is also obvious, and not really applicable here. The client doesn't care much about that one way or the other, at least the tech guy doesn't. They will be running on Solaris and Oracle for sure, so they are quite comfortable with proprietary platforms.

This is a straight technical discussion not a philosophical one (although the *nix vs MS philosophy is a very important discussion, so feel free to pontificate on that one.)

Some of my questions include:

• What does .net have that OACS doesn't? Will this be *forever* true?
• What does OACS have that .net doesn't? Will this be *forever* true?
• How do RAD times differ?
• How do the developer communities differ?
• How do learning curves differ?
• Any other questions or statements?

I know of at least a few people who are intimately familiar with both application environments. We have certainly exhausted the Java vs OACS discussion, which probably should have been framed within J2EE vs OACS.

TIA

talli

Based from low experience with .Net here is my best shot.  This is just some facts I know, not advocacy of OACS or .net

*  What does .net have that OACS doesn't? Will this be forever* true?

Its not going to hold forever true, but it might.  .Net advantage is having an IDE (Visual Studio .Net) that really makes it easy for fist time developers to start working on .net.  Also those who are coming from older Visual Studio will be familiar with things.  Not that its is the same, its different.  But the perception is VS .Net is the upgrade.  OACS have emacs, which may not be acceptable for newbies.  We are unlikely to ride on Eclipse.

* What does OACS have that .net doesn't? Will this be forever* true?

License, we are Open Source.  The ability to run on non MS platform.  Also we are more proven technology.  ACS goes back a few more years in web technology compared to .net.

* How do RAD times differ?

Starting from scratch will have .net faster, care of more tools and VS .net.  Once developers are familiar with the platform OACS should be faster.  As you are free to go deeper into the guts.  Also tcl and sql are very easy and forgiving languages.

* How do the developer communities differ?

We are smaller, .net community is bigger.  OACS has only 1 community, while .net will have different types of community.

* How do learning curves differ?

Depends on where the user is coming from.  SQL junkie will be right at home with OACS.  Non SQL guy will have a harder time with OACS.  Also a VS experienced developer will have lower learning curve to .net.

* Any other questions or statements?

OACS problem space is on web applications, although some have done work beyond this.  But its real niche is web applications.  While .net being a marketing umbrella is more than creating web applications.

One would have to know a lot about the particulars of their environment to determine which is the better choice.

Technology choices need to be viewed as business decisions and one of the key business questions is sustainability. It's one thing to build an application as a one-off. It's quite another to keep it evolving in response to user needs and new markets.

If I were in their shoes, some of the questions I would ask: How much does this application need to evolve? Do I have .NET expertise in house and the budget to continue to enhance the application?

OpenACS has an inherent advantage over .NET and that's the *community* of very smart people supporting and developing the toolkit. For us SloanSpace/.LRN has been a huge success. Could we have done it in J2EE or .NET? Sure. It would have cost us a hell of a lot more.  More importantly, we would not be in a position to leverage the ongoing work and innovation of the community. That's the beauty of open source.

A good technical argument would be, OACS has lower hardware needs.  That is why OACS appeals to smaller organization that is tight on a budget.  You can say outright there, take this desktop machine of yours.  And lets install OACS.  You can develop, use it for live site.  Take the same machine and it will hardly be usable for .net both in serving for the live site and development.  So in .net you need to buy your live server and development workstations.  But then hardware is cheap this days... so not a real $ saver.  Just a perception.

You can take the argument further.  By saying that lets get all the junk PC in your inventory and make them into a cluster.  To serve that kind of traffic .net will need X amount of hardware.

Another is inverse the argument of .net.  As I have stated above, .net is easier because it has VS.net.  You can say, .net needs a IDE to develop, so its harder.  So additional cost in purchasing the tool.  While in OACS a text editor is all you need.  Then give them a very short demo and get their hands dirty on a template.  Show them how easy to use the template or even ad_page_contract.  Or even a very simple ad_form.  Although such functionality can be done in .net, you will need VS .net to ease the pain.

You can also show them that .net's web form are generated from a DLL that MS ships.  This would mean its not technically possible to alter it.  Unlike in OACS the forms source code are all there.  Its with templating's form builder and given to the developer in easier fashion in ad_form.

Still with all the argument that you can come up, Al is 100%.  It will depend on the business context.  You are unlikely to walk out from a MS shop with a signed contract.  Also it maybe likely possible, that all your arguments will be subconsciously telling them that they are dumb in deciding to invest in MS.  :)

1. lock-in
2. ever-increasingly onerous licensing terms including near-mandatory updates, etc
3. reliance on X86 hardware
4. licence audits more akin to a car-jacking than anything previously seen in the software world.
5. Mono?  MS is patenting key components of .NET
6. security.  Pointing out their history isn't FUD.  Yes, they've claimed to be paying more attention but there've been a couple of nasty rounds of exploits since and no sign they're letting up.
7. If you want to live in the X86 space, 2K requires more resources than a headless Linux server box, and Linux wins hands-down in reliability.

There's a *very* good business case for staying as far away from MS as possible.  Hear those footsteps running down the corridor?  Those are IT managers running towards Linux-based solutions offered by real-world companies like IBM.

OpenACS - and some other potential solutions - really allows you to scale your solution to your problem.  Are you large and feel you need the world's most popular enterprise-class RDBMS and want the security of running it on the most popular hardware platform for hosting it in the enterprise?  We support Oracle and Solaris, and use the same webserver as the busiest web site in the world, aol.com.  We also support vanity blog sites running Postgres on cast-off PCs.

I've used Visual Studio professionally back in my compiler-writing days, as recently as four years ago.  It's a decent development environment and we don't offer anything like it.  I'm sure that it's just as great for writing .NET apps in C# or VB as it is for C++.  I'm sure that .NET's wide set of classes help one cobble together apps quickly.

As far as XML being as important as the PC or the spinning electron or whatever, remember that XML - like HTML - is simply a subset of SGML, which has been around for a very long time ... that's hype and nothing but hype and even if it weren't hype, MS isn't alone in using XML.  You could point out that OpenOffice and KDE have used publicly-defined XML formats to store documents for years, now, since before Steve Ballmer learned how to pronounce XML.

I think there are plenty of ways you can attack an MS solution without referring once to OpenACS ... which then becomes a bit like the icing on the cake.

But if they're sold on MS then they're sold on MS and your sales budget is probably better spent on the increasing number of people who are seriously considering non-MS server solutions *without* being prodded by a vendor.

But ... the organization seems screwed up anyway.  Board members shouldn't be making operating decisions of this sort unless they happen to be part of the implementation team.  And that would be unusual in most organizations.  And he should be clear about which roles he plays - if he's the project lead then his ass should be on the line for completion.  If he's acting as a board member he should trust his staff and allow them responsibility because it will be their ass on the line.

My guess is that his ass isn't on the line but he's going to do his best to make those whose asses are on the line do things his way regardless.

The other issue is that in essence he's saying he doesn't trust Musea, a potential vendor, to make an intelligent choice.  That also raises red flags IMO.

If it were me, unless I really, really cared about the organization I'd run away and never look back ...

If you win the business and do a good job, your board member will probably continue to say "it would've been better in .NET"

If you win it and screw up, due to faults of your own, the inability of the client to spec a system that meets their needs, or both, you'll here "the project would've been a success in .NET"

What you're *never* going to hear from this board member is that they made the right choice by going with Musea and OpenACS.

You know that campy saying some people use ... "What would Jesus do?"?  Ask yourself what Bill, the antichrist, would do.

When's the next board election?  How much would it cost to oust this guy?  That's what MS would do! :)

http://www.ibuyspyportal.com/DesktopDefault.aspx?tabindex=4&tabid=5

And this was more than a year ago.

Back when Enhydra was a viable contender in the open-source J2EE space, I asked one of the project heads why they don't provide any applications or data model.  He said that this was a conscious decision not to interfere with the vendors of add-on applications.

I suspect that Microsoft has a similar viewpoint.  Namely, if they provide a data model as part of .NET, then (a) it won't be appropriate for everyone, and (b) it'll mean that Microsoft is competing with their own integrators and vendors.

So while .NET has all sorts of fancy database-access methods and marshalling systems, I expect that you're stuck doing all of the database work on your own.  This is where OpenACS has a big win, for sure.

talli