Tim O’Reilly says ASP.NET is on a roll

ASP is fine, but Ruby on Rails (and AJAX on Rails) pretty much offer the same scripting capabilities, without all the strings attached.

ASP makes me think of IIS, and thinking of IIS makes me vomit a little.

2.0 really is all it’s cracked up to be. App-specific code (I still have to augment the web controls with my own library to fill in some databinding gaps) is rarely required, meaning most of the app can exist in the .aspx instead of the code behind. It’s a real joy using it instead of 1.1.
Too bad VS2005 is still a bit flaky/unstable, though (and the debug-time file locking is a step backwards). But then, the 2003 IDE was worse in a lot of ways, particularly with assembly locking & project caching.

2.0 really is all it’s cracked up to be. App-specific code (I still have to augment the web controls with my own library to fill in some databinding gaps) is rarely required, meaning most of the app can exist in the .aspx instead of the code behind. It’s a real joy using it instead of 1.1.
Too bad VS2005 is still a bit flaky/unstable, though (and the debug-time file locking is a step backwards). But then, the 2003 IDE was worse in a lot of ways, particularly with assembly locking & project caching.

tetra and Tim, have you honestly tried IIS lately? Sounds like you’re still thinking of the older version. The new version runs Canada’s biggest Google AdSense site and the guy who runs that says it’s awesome (and is dealing with very high loads).

Yeah, I’ve tried IIS lately. Its fine, no real complaints. Its just that Apache is superior in almost every regard and is extensible in ways that IIS just can’t match. Versatility is the name of the game where I work, and IIS requires an overall investment of time and resources that its capabilities just don’t warrant.

yes, ASP .NET is very good for web app, but unfortunately your blog is powered by PHP and not ASP .NET so you have no credibility to say anything about it! sorry.

Admittedly, I haven’t used IIS lately, I’m an ex-microsoft developer who was one of the first to leave .NET as soon as I saw RubyonRails, but it recently came to my attention that Rails and .NET don’t really compete in the same arena, comparing them is like apples and oranges. They are two different animals for two different arenas. .NET will always dwarf Rails in terms of the number of developers who use .NET and it will be the big commercial force for web development that it should be. Rails will continue to attract a disproportional amount of press as it’s going to continue to gain the most talented developers into it’s ranks.

I’ve proposed Rails for some .NET shops and then reconsidered if Rails makes sense for them. Some businesses are built on the model of doing very regular work for regular clients. If a .NET shop is filled with beginning to medium level developers it’s highly unlikely that they are going to have the motivation and stamina to get over the learning curve that is Rails, even if they are promised from the beginning that they will become better, more productive developers, and some businesses simply aren’t based on having the most talented people, some are based on what’s economically advantageous, and that means if you can get .NET talent cheaper then you can compete on cost. If you compete on cost why hire expensive talent?

The Rails developer though will still be able to do the job of three to five (maybe even more) average .NET developers, but the Rails developers won’t be doing the same jobs.

.NET has it’s place where it does well for a particular situation.

To get back to IIS, I haven’t used it lately but I still keep in contact with those who do and we have conversations like this:

ME: So what are you doing today?

.NET GUY: Deploying

ME: All day?

.NET GUY: I hope it’s just today, but you know how things go with ASP.NET and IIS, what are you doing today?

ME: I’m doing another release as well.

.NET GUY: All day?

ME: Nope, just finished.

Here’s my experience with deploying an ASP.Net app to IIS:
1. copy files from build directory to virtual directory
2. see if the page is there
3. If the page is not coming up, push Stop button. Push play button.
Takes 30 seconds *tops* (and the IIS restart is a couple of seconds). For comparison, doing the same thing with a comparable WebLogic J2EE app takes about 10 minutes, as the launch process is orders of magnitude longer in duration.

Well, to each their own, but I find Apache to be a typical Unix-spawned package, as cryptic as Hell’s Ninth Ring is horrible. Of course it functions. That’s not in question as it’s a prerequiste for this type of software. The problem is that Unix-style server code sanctions and supports making things as unintuitive as possible. This makes it easier for admins to act like prima donnas, because it’s so difficult for mere mortals to understand and remember how to use admin tools.

IIS has any number of flaws, but at least any developer with a fundamental understanding of how Microsoft does things can admin IIS without needing a week of class.

Comic: get lost. Our blogs over at http://blogs.msdn.com (which has more traffic, and thousands of bloggers than me) is running on ASP.NET.

My video blog, http://channel9.msdn.com is running on .NET.

My wife’s blog, which is running on MSN Spaces, is running on .NET.

So?

Tim: sounds like you have some incompetent friends.

We deploy regularly on ASP.NET and spend very little time doing so. (Our video blog/community is on ASP.NET and serves 2.8 million unique visitors per month).

And wait until you see what is coming next from our team. We developed it all in less than a month and it kicks ass (sorry, can’t show you it yet, but it’ll be all over the place within days).

Oh, and go and talk to the http://preview.local.live.com/ — that was built in less than a month too. All in .NET.

“Its just that Apache is superior in almost every regard and is extensible in ways that IIS just can’t match.”

Please, enlighten us.

“We deploy regularly on ASP.NET and spend very little time doing so.”

Yeah, since Microsoft has such a great reputation with their deployment times. Hyuck. How long did it take you to ape Google Maps again?

Tetra: that had nothing to do with deployment times. Thanks for proving, once again, that you are a religious advocate without any understanding of software development processes. Why hasn’t Google matched A9′s and our “street side” views yet? Hint: it has nothing to do with deployment times.

Scoble: I retract my previous statement in which I implied that a .NET deployment takes a day, that was my experience a couple years ago at the company where I worked, I don’t know why but the people who did deployments took a whole weekend. I spoke with one of my .NET friends this afternoon and he says that it only takes about 25 to 30 minutes for them to deploy, and that he and his company are very pleased with .NET.

His only complaint of .NET is the lack of sufficient reporting integrated into visual studio. They tried SQL Reports and found it lacking to say the least. Still they love .NET and planning a new year of growth and hiring for it.

Wow, I guess I’m one of the few “advanced” developers who actually prefers ASP.NET to Rails, and I’m quite familiar with both technologies. Ruby is my favorite programming language, but the characterizations I’m hearing that RoR is for advanced developers while ASP.NET is for beginning to intermediate ones is hogwash. As a matter of fact, I find that Rails works best when a database is viewed as merely an object serialization repository and nothing more. Bye-bye relational model. (Auto-generated identifiers used as primary keys and required by ActiveRecord aren’t relational. Sorry.)

That’s sometimes an OK approach, but I’m one of those database purists that the inventors of Rails rail against. I actually believe that the relational model is the best one for representing data, especially when the tables you create will be used by more than one application and you plan to keep scaling up.

If the tables you create will be owned only be the RoR application, and the app has a small, targeted focus, I think it’s a great little platform. But it isn’t suitable for the big stuff, and if you think it’s naive database schemas play nice with others, you’ve probably never heard of Edgar F. Codd.

Oh, for the love of God almighty, will someone PLEASE explain why “Apache is superior in almost every regard and is extensible in ways that IIS just can’t match”.

Anyone?

interesting.. but i’ve noticed that the smartest, most efficient programmers I know (working on a large range of products) don’t buy many O’Reilly books.. all of these languages provide endless possibilities and scalability, and it seems documentation is typically found online these days. I never found O’Reilly books to be a useful source for learning new languages.. they’re best for reference.

I like RoR too, but I haven’t seen any real enterprise web app built using RoR. J2EE or .NET is still dominating most enterprise application developments.

ASP.NET as part of .NET has its place and is yet another choice in server-side scripting solutions. Where I feel developers and vendors really fail is in addressing which technology is best suited for a given responsibility within an architecture and which technology is best when taking in non-technical factors.

Recently I had a horrible interview with a company that previously had outsourced all of its development, was admittedly short on technical resources, and somehow had managed to be sold on the idea that version 2.0 of their product *must* be .NET based. When I asked why they were moving forward with .NET they couldn’t answer! The reason they couldn’t answer is because .NET was sold to them as a magic pill solution for their product. The product? Templatized web sites with very few pages for individual financial advisors similar to any $9.99 php/mysql based hosting plan – something I could have whipped up in 3 weeks customizing Drupal or some other CMS. BTW I should mention at this point that they had just failed over the last 6 months to deliver an ASP.NET solution working with two separate consulting firms.

Upon leaving the interview I made it very clear why I strongly suggested their ten person non-technical office choose PHP/MySQL if good .NET reasons couldn’t be found:

- The extreme difficulty of them finding a senior enough .NET developer that would work for their offered $60k/yr. compensation VS. finding a fresh young and excited PHP/MySQL for the same rate.

- The ability to be able to host in a cost effective manner with a degree of migration flexibility between platforms.

- The ability to engage many more consumer internet focused web designers who make things look pretty. More artsy types can be found who can work with xhtml/css/php/mysql as a technology stack versus dealing with designing around ASP.NET code.

As follow up, where I feel .NET shines is in the area of needing to create enterprise applications, xml web service based applications, and applications for which an outside content provider has supplied a very specific interface that lends itself toward consumption or interaction with a .NET client. But quite honestly, if my retired father or some other non-geek friend were to ask me for a hosting provider and a blog page, ASP.NET would not be the technology I would suggest. In fact with javascript/ajax supplying “good enough for CRUD” type interactivity I find myself suggesting ASP.NET even less. For someone wanting a few pages and some pretty design and simple transitions, why would I ever suggest something as cumbersome as .NET!?

There are also many companies that are migrating .NET apps to Rails.

The one thing that will never change is….change

Look at google!

asp is fun and exciting and brings knowlege of internet programming fun but things such as java and c comes into place and may over whelm the programming techniques used by most asp.net users.

I have found it very difficult to get the knowledge on board quickly to adapt to the market. By time you have found all the quirks in one programming system another is starting to emerge. The best advice is to learn one method and stick to it and just buy into the fact it wont cover everything.

Andy