Debate - .NET V. PHP: Top 6 Reasons to Use .NET

[mgkimsal]

Originally posted by Jeremy W.
The post was in response to Harry's "how would .NET react to [x]".

We didn't redo the TimeSheet application in 10 minutes, we had a production level upgrade ready to go and simply hadn't deployed it yet (since it would have meant taking the service down for a few minutes, and we were going to do it at night).

I'm not knocking PHP or even saying "can PHP do THIS", just answering questions

I know you're not. I was just wanting clarification on the '10 minute' thing. Yes, installing a package in 10 minutes seems reasonable. I had initially read it as you *redid* the timesheet app (after it crashed) in 10 minutes. If that was actually possible (write a full app from scratch in 10 minutes) even *I* might be tempted to switch.

[mgkimsal]

Originally posted by Mincer


I'm sorry, but I fail to see how writing all your code in the way that MS wishes is such a good thing? Surely choosing the library that works the way that YOU want to, is infinitely better than changing the way you work (even if you KNOW that it's not the best way to solve a particular problem), because you want to work with one single library.

You prefer to do things wrong because it's the way *you* want to work? What about value for money for your clients or employer (assuming you have clients or a PHP job and aren't doing this for a hobby)?

Make arguments about the quality of the .NET libraries or Java libraries or PHP libraries, but don't argue with me that doing something subpar WHEN YOU KNOW IT IS SUBPAR is somehow more noble because you have a 'choice' to do so.

And if we're bringing security issues into it, knowing the speed at which MS addresses internet security problems (SSL??), is it a good idea that every .NET website will have the same security issues when they are uncovered.

It's about as good an idea as having 25 PHP books in publication (which will be circulating for YEARS) demonstrating utter crap code examples with respect to scalability and security.

Systems can eventually be patched - unlearning years of bad habits reinforced through article after article and book after book across thousands of developers takes considerably longer.

[Mincer]

Originally posted by mgkimsal
You prefer to do things wrong because it's the way *you* want to work?

But that's based on the assumption that the .NET librabies are all written in the 'right' way. What is the 'right' way btw?

Originally posted by mgkimsal
It's about as good an idea as having 25 PHP books in publication (which will be circulating for YEARS) demonstrating utter crap code examples with respect to scalability and security.

Systems can eventually be patched - unlearning years of bad habits reinforced through article after article and book after book across thousands of developers takes considerably longer.

So, are you saying that using .NET, or writing books/tutorials for .NET suddenly makes you a better programmer than you were before? Surely the fact that there are multiple language.NET's means that the scope for writing horrific code is multiplied? *shrug*

Personally I think that .NET is a very good idea. My only worry is the fact that it's being touted as the solve-all solution to any problem.

[mgkimsal]

Originally posted by Mincer


But that's based on the assumption that the .NET librabies are all written in the 'right' way. What is the 'right' way btw?

So, are you saying that using .NET, or writing books/tutorials for .NET suddenly makes you a better programmer than you were before? Surely the fact that there are multiple language.NET's means that the scope for writing horrific code is multiplied? *shrug*

Personally I think that .NET is a very good idea. My only worry is the fact that it's being touted as the solve-all solution to any problem.

*YOU* suggested that you'd prefer to use libraries that work the way you want, even if they were 'wrong'. You brought up the idea of 'right' and 'wrong', not me.

Never said it makes you better, but having a community with one common way of doing something has benefits which aren't easily seen by people who continually want to 'do their own thing'. Both .NET and Java have a more consistently unified approach to low-level handling, which is adhered to by the majority of their respective development community. The same can not be said of PHP, unfortunately. PHP might well be in use by more people (now or in the future) but they won't have common codebases to build from. Some people are worried about being PEAR compliant, some working on PostNuke 'hacks', etc. We're the same way- not saying we have the answers, but I'm not sure most people even agree there's a problem.

[Mincer]

Originally posted by mgkimsal
*YOU* suggested that you'd prefer to use libraries that work the way you want, even if they were 'wrong'. You brought up the idea of 'right' and 'wrong', not me.

I never said anything about right or wrong. I said, and I quote:

Surely choosing the library that works the way that YOU want to, is infinitely better than changing the way you work (even if you KNOW that it's not the best way to solve a particular problem)...

If I find some quirk/bug/blatent error in the .NET libraries, am I supposed to use them anyway?

I think these debates are showing more and more than there is no 'winner' in any of the catagories put forward for deliberation. Whether you want to use any one solution or another, is really up to you.

[Jeremy W.]

I think the debate is fine, it's just gotten rather pedantic. You're trying to argue something you don't believe in, against something you know little about, so it's coming across as you and mgkimsal just going at it over nothing.

Are the .NET class libraries bad? You don't know.
Is the "MS way" (whatever that is) bad? You don't know.
Are there defficiencies in the way .NET handles large project volume? You don't know.

Stick to what we know and we'll all be better for it

[mgkimsal]

My fault - I misread 'better' as 'right'.

[Mincer]

Originally posted by Jeremy W.
Are the .NET class libraries bad? You don't know.
Is the "MS way" (whatever that is) bad? You don't know.

My assumptions were that it is, on the whole, very good. But I was also assuming that it was not perfect (since when has anything been perfect?). If there is a definative methodology to problem solving when programming, then why is it that I don't know about it? And if there is, but in my ignorance I've missed it, can someone point me in teh right direction? There's more than one way to skin a cat.

[Jeremy W.]

Of course there's more than one way to skin a cat, but you are trying too hard to play the devil's advocate. "What if .NET's class libraries aren't all they're cut up to be?"

What kind of argument is that?

If you have a real gripe with .NET, we can answer it or say "you're right", but to say "what if" just leaves me going "well, what if PHP only let you type in Spanish?".

[mgkimsal]

Originally posted by Jeremy W.
I think the debate is fine, it's just gotten rather pedantic. You're trying to argue something you don't believe in, against something you know little about, so it's coming across as you and mgkimsal just going at it over nothing.

Are the .NET class libraries bad? You don't know.
Is the "MS way" (whatever that is) bad? You don't know.
Are there defficiencies in the way .NET handles large project volume? You don't know.

Stick to what we know and we'll all be better for it

Didn't it was getting pedantic.

'bad' is relative.

I was trying to point out that - as much as I'm not a .NET fan, I see value in the standardization MS developers have (and Java too). PHP is, for the most part, still a 'wild west' area, relatively speaking. Some of this is due to the open source nature - people prefer 10 ways of doing something in the name of 'choice' over having one way which everyone pretty much accepts, arguments over the quality of the 'ways' put aside.

[Jeremy W.]

I was generally talking to Mincer and his devil's advocate posts

[Mincer]

Originally posted by Jeremy W.
Of course there's more than one way to skin a cat, but you are trying too hard to play the devil's advocate. "What if .NET's class libraries aren't all they're cut up to be?"

What kind of argument is that?

It's the same sort of arguement as saying. Hey, look at this new .NET technology with it's full class library framework. It's got one big framework written by MS, so it must be good. That's not nessessarily aimed at you Jeremy.

What exactly are they "cut up to be"? Surely a framework that's only been available for a few months can't have been fully evaluated yet? I was merely exercising caution, rather than jumping head first into the fire bacuase someone told be the t-shirt I was wearing was flameproof..

[Jeremy W.]

Well,

Thing is they have been evaluated. Thoroughly even. I mean, .NET's been around for over 2 years in some form or another. We've been using it on Windows Media Server for over a year (production level, largest stream in our industry).

We've been developing in it for nearly 6 months. We've laid down hundreds of thousands of lines of code, and dumped out some pretty kicking projects.

Others are even farther along. No, it isn't "as tested" as other technologies, but that's like applying for a tech job these days. There will always be some technology that's older and more stable.

Evaluate .NET based on what you know, ask questions about what you don't. Before you were making statements like "using .NET's class libraries is dangerous" and so on, without knowing.

I understand caution. Caution's a good thing. Nobody is "preaching .NET" in here at all, we're simply (if you read through the thread) answering questions and fending off attacks, as well as getting rid of misconceptions.

In reality, several people have had their eyes opened. In fact, don't tell him I told you this, but Sketch was in my office yesterday and we were giving him a tour and he's nearly converted to the possibility of learning .NET.

Similarly, Harry, as far as I know, is using 2 of our .NET apps. Password Generator and ActiveCS. Together they are NSA-level encryption, and possibly the best and strongest standard-algo security you'll find.

So, while some people are arguing, the reality is that others are actually using these technologies. Personally, I like that everything I need is at my fingertips. For every feature .NET has, there may well be a PHP equivalent, but is THAT tested, or is it just there to copy? I mean, who wants to search through the 50+ links Harry posts, just to compare to 50 class libraries in .NET. There are 15,000 classes in .NET, I can't imagine having to download even 1% of those, and keep up with updates, fix security patches, etc.

[Mincer]

Originally posted by Jeremy W.
Evaluate .NET based on what you know, ask questions about what you don't. Before you were making statements like "using .NET's class libraries is dangerous" and so on, without knowing.

You have convinced me, in the last few weeks, to have a much closer look at what .NET can offer.

And I don't think I've ever made the sweeping statement that using the .NET classes is dangerous. Perhaps it's you who just _wishes_ me to be devils advoate?

[Jeremy W.]

Maybe, but since I haven't really seen you outside this thread, why would I?

Anyways, as I said, let's move on.

[HarryF]

My point with "doing things the MS way" is that's what an application framework does. Yes I know there are many solutions to problems and you can write your won classes, but to be successful using an application framework, you should follow it's guidelines for development.

If you look at the IBuySpy examples on the asp.net site, they are provided mainly as examples of how to develop web interfaces with ASP.NET.

So here's my point: new technologies come along and everyone says "This is great - walk this way!". And everyone does, investing alot of time, money and energy. Then someone else realises there's a better way to do things and for some of the adopters of the earlier technology, they've actually wasted a whole load because their technology is now redundant.

An example from the MS world is DCOM (and perhaps even COM+) - how many legacy applications will need to be rewritten to interoperate with the latest .NET apps?

Another example this time from the PHP world is templates. Everyones been raving about templates being a great thing, claiming they seperate content from logic (which they don't) and also that they make web designers lives easier. Does this look easy to you?

Many PHP developers thought they were actually building scalable, N-Tier applications by using templates. In fact they invented a whole load of template languages which can only be used in one environment and a whole load of PHP apps that don't fit well together (try adding forums to your site which already has a user login system, without ending up with two login systems).

This is where I raise questions about .NET. First of all, web application building is still a young discipline. No one has the perfect answers yet and even big companies like Sun and Microsoft are still trying to work it out.

I need to look for myself at what IBuySpy encourages but some of the things I've seen with .NET don't look to me like true layering within an application. Microsoft may be taking many people down a road of web application design that doesn't work well, much like PHP's templating mania.

And let me just add one more example, using Jeremy's application. You have an authentication system for Services. Now that seems to imply the application will interoperate with other companies applications, using SOAP?

Let's say Microsoft chooses a particular approach for security within SOAP, build it into the .NET class library and Visual Studio. Then someone discovers later (one of the Vincents of this world) that appoach is fundamentally flawed. The companies you want to interoperate with adopt a different standard. What investment is required to convert your existing .NET applications? (I mean that as an open ended question - right now there is no answer).

[Jeremy W.]

I agree, just to shed some light on your question about how we do things:

Our Auth system doesn't use SOAP at all, it uses Windows Remoting, which is basically an Active TCP/IP connection (same kind of thing as when you request a file from the server).

It's encrypted, it's secured, and it's IP to IP, port to port (dynamically assigned).

So, no, we aren't providing extranet functionality, even though the web services we use, and the WR stuff we write, is on the web server. None of it is externally accessible, and there is no reason for it to be.

We expose interfaces and so on to ASP.NET pages for our users, but again that's the ASP.NET calling it (IP to IP, port to port, authenticated user accessing a sandboxed app).

So, sure, we are kind of doing things the "MS way", but you can't on one hand say MS is doing nothing new (TCP/IP connections certainly aren't new) and then say that the "MS way" is so vastly different from anything existing

[Technosailor]

Originally posted by Jeremy W.
In reality, several people have had their eyes opened. In fact, don't tell him I told you this, but Sketch was in my office yesterday and we were giving him a tour and he's nearly converted to the possibility of learning .NET.

What now?

[Sillysoft]

Thought I give my 2 cents. I read a couple of pages but didnt feel like reading the rest

.NET development is free but you still have to pay for the OS. WebMatrix is free, the watered down version of SQL is free and .net itself is free. Is .net a language or just a "house" to use other languages? Or both?

Why does it take so many lines to access the db? Though it is easy because WebMatrix has a GUI interface with a wizard to do all the quering and db connections for you.

But is that really learning anything except how to use WebMatrix?

Silly

[Jeremy W.]

.NET is a framework, not a language. That said, the languages within .NET are nothing more than syntax, generally speaking. So, .NET is the "power" (foundation/base classes), the structure (how to access what) and the framework.

Do you really want to get into "does the GUI actually teach you anything"? You'll have some heavy cards to play when you start talking to app developers who would never write apps from within a text editor, even if they wanted to.

Why so many lines to make a DB connection? It's just 4, roughly the same as most folk use in PHP. Besides, number of lines makes no difference to speed in .NET, and only makes a difference if you're handcoding, which I've never seen anyone advocate, besides for learning.

Why is there so many lines, though? Because it's object oriented and you need to instantiate everything properly

If you don't want to use Windows to use .NET, use Linxu with Mono. Done, no licensing fees at all. And then you can use it on Apache with MySQL if you want

J

[Sillysoft]

I never said either way which language I like. Just my opinion from my experience.

Silly

[Jeremy W.]

I know you didn't say which language you liked I don't even know which language you code in, if any. I was simply saying that it's not really that many lines of code for a DB connection. PHP, ASP, JSP, CFM all have 2-5 lines of code, and .NET isn't much different

J

[Sillysoft]

Yep

Yes I have used Webmatrix and the experience in my opinion is I felt I was missing a learning curve atleast for connecting and querying a db as well as connecting to it. But Im sure I will get over it as I get more into .net

I code primary in php but am making the move over to c# with .net

Silly

[M. Johansson]

To the database access part, I just have to say: Huh???? I don't know what wierd moon version of .NET you people are using, but in mine, it doesn't take much code at all:

//Connecting: One line
SqlConnection myConnection = new SqlConnection("The Connection String"); 

//getting a dataset: 3 lines
SqlDataAdapter myCommand =  
new SqlDataAdapter("SELECT * FROM Guitars", myConnection); 
DataSet ds = new DataSet(); 
myCommand.Fill(ds, "Guitars");

//Displaying the data in a table: 2 lines!!!
GuitarSelectBox.DataSource = ds.Tables["Guitars"].DefaultView; 
GuitarSelectBox.DataBind();

[Sillysoft]

All I can say is I might be used to using classes where I dont see that code, only the functions that run it. Which is a far less to the eye compared to seeing the actual code in action.

Silly