Why web programming?

[brandonb]

This is related to the other thread just posted recently:

http://forums.anandtech.com/showthread.php?t=2285528

But my question is the reverse. If I've been nothing but an application developer for 20 years, why would I ever want to work with web programming? Alot of people in the thread above have said "Stay web programming." The question is why?

Every time I've tried it, be it Ruby, or ASP.Net, I found it extremly sloppy all around and had a huge dislike for it. The entire time I'm working on these projects I'm thinking to myself "If I could just make an App this would be so much quicker to build." I also could not handle the lack of a debugger found in Visual Studio. Things just seem like a patchwork of technology and it only works due to miracle.

I can see if you want to create an app on a webpage for browser users. But for internal business apps I don't see why I'd want to use the tech.

Take web enabled apps, such as uTorrent, Microsoft/Yahoo Messenger, etc. A web based solution (like the Facebook chatting app) just doesn't feel the same as using a native desktop app.

What is the benefit for web programming for an application developer? Try not to say things "It's the way the industry is going..." That is not a reason. The industry is only going that way because people are building web apps. But why are they building web apps to begin with? I recently had someone at my job (who knows nothing of technology or programming) wanted me to make a web page for doing this internal app. I asked "Why does it need to be a webpage?" He responds "Oh no reason. If you can do it another way go ahead."

Is it because it's a security risk to install apps? The maintaining versions and updating your machines with the new software?

---

Please keep it civil. This is not a bash on web development. It's just not for me, and I will retire before I make web apps for a living. But I'm curious what the draw is for others because I've never seen the appeal.

 

[Markbnj]

It always seems to me that these questions spring from a fear of being pigeonholed into the "wrong" discipline. I guess that's not simply paranoia, which is sad. When I started everything was written in C, assembly, BASIC, or Pascal, and had a character mode interface (or even just line printed output). In just the past four years I've worked on WPF, Silverlight, iOS, ASP.NET, MVC3, MVC4, C++/Win32. Last few months have been all MVC3 and javascript with some html/css tossed in. I don't think of myself as a "web programmer." I'm a programmer. I have Google, and if you can describe what you want I can write it. I refuse to be typecast.

 

[gregulator]

Yes, moving from a controlled platform is tough. Web programming relies on a ton of pieces working together, and many of them don't have the maturity of a language like C. Not to mention that the target platform can be a zillion different devices/combinations.

I am with you though brandonb, whenever I have to do any web programming, and try to just put a picture in the middle of a page, I cry a little on the inside trying to figure out how to make it look the same on all platforms...

But businesses like the idea of write once, run anywhere. Just like the promise of Java. But like Java, it is so difficult to make that a reality. It is more like, write once, then make a bunch of exceptions and requirements for different platforms.

 

[Ken g6]

I should preface this with the fact that I've only managed to develop GUI applications for three platforms: Visual Basic 3.0, Java, and web apps. A few things I've noticed about web apps:

- Their development tools are free in most cases. Maybe this isn't quite so much of an issue these days, but before MS started giving out free Visual Studio Express Editions, practically all GUI development tools cost money. Often quite a bit. Java was one exception, but Java has its own issues on the desktop.

- They're cross-platform. Maybe not quite as easily as Java, but you need a JRE for your environment. Way more easily than C++, or C#. Can you run a .NET app on Linux? Maybe, with Wine or Mono. What about on an iPad? Not a chance! Of course, there are advantages to making one app per platform - just ask Facebook.

- From a business point of view, you can discreetly collect all sorts of data on your users. Just ask Google. A desktop app could do that too, but people would start asking questions about why it has to connect to the Internet.

 

[PhatoseAlpha]

There are also other advantages. A web-based app exists in one place, one place only, and is always that particular revision. Which means you don't run into issues where you're dealing with bugs fixed 4 revisions ago because users don't update.

Plus, there is the mobility aspect of it. Smartphones. Users want to be able to access their applications on the go. Which means you're either writing a web app, or you're supporting Android/iOS versions, possibly WinPhone/Win8 or god knows what else. And given that Apple has a pretty notable presence in that space, that means you're realistically dealing with their approval process for your app - and nobody likes another company having a blanket, no reason required veto over their apps.

It all adds up to "Your choice is to support 3-5 native apps, some of which require outside approval" or "Write in once in HTML". The native apps have advantages, sure - but are those advantages really worth the costs?

 

[Markbnj]

I am with you though brandonb, whenever I have to do any web programming, and try to just put a picture in the middle of a page, I cry a little on the inside trying to figure out how to make it look the same on all platforms...

Just recently I was able to, for the first time, work with a top-shelf front-end designer and markup guy who produced bullet-proof html/css for a client site. All I had to do was turn it into cshtml views and the result was consistent across all browsers. It was awesome.

 

[brandonb]

I'm responding to my own post, someone feel free to enlighten me.

I look at a web browser as the view or UI to a much larger app. Sure, you can write HTML and if you are lucky, write the css etc to make it look good across all browsers. But you really are at the mercy of the web browser being standard, compliant, and you have the headaches of that. Let's discard that and assume every browser renders the same.

I understand the benefits of the web browsers creating a standard front end and there is only one app to install, the web browser. Sounds great on paper, and I understand the benefit to that.

Now we need more to web apps than just the view/UI. What if you want to read/write to a database? Now you have to go through the web server. Using different frameworks here and there.

I believe the simple thing that turns me off to web programming is HTTP servers don't seem to hold state very well. It's stateless for the most part. With load balancing and server farms. Everything has to be stateless. Which means you have to jump through hoops to create an app that needs to be stateful. So you spend alot of time on the server side to accomodate for a simplier client side.

The simpliest scenario in web programming that I have trouble with, is that the client needs to query the server for information in most if not all cases. The polling technique. The server can't push information to the client. Which means that its role is limited to me. The only way to get decent response is if you poll frequently and with the web based app all those users/browsers polling takes a toll on the servers.

Am I missing anything or has the technology matured past that?

 

[Markbnj]

Am I missing anything or has the technology matured past that?

Not really. It's all request/response in the http world. Mobile apps can do push notifications, but most of what they do is still request/response. The web was originally designed for resource retrieval, and the last nearly twenty years have been an exercise in pushing it well beyond its original design goals. The reason it has been pushed, and continues to be pushed, beyond its own limitations is all of the other benefits already cited.

I'm not sure I understand your point, ultimately. Technology stacks are good for what they're good for. If you were asked to create an application that lets the public view upcoming movie times at a theater you wouldn't write a client-server Windows application. The web would be the obviously correct choice. Thinking of examples for the inverse case is harder, but there are some things that native applications still do better. Audio engineering, for example, or CAD, or whatever.

Are you really saying that you don't find web applications interesting, and wouldn't want to learn to write one for that reason?

 

[brandonb]

What I'm saying is as an application developer, I'm trying to find a good reason to learn web programming as well as I know application development. If I'm creating something that is external, to the consumer, such as viewing upcoming movie times at a theater, sure web page/web app all the way.

But typically in every place I've ever worked, they typically sell software to places such as brokerage firms, banks, remit/billing departments, etc. And web development seems like it would be a waste of time to explore. Unless you wish to expose your app to the outside world for the masses such as online banking.

I'm working on collection software at my place of employment, and I was told by the IT staff I need to make it web based (its the future!). But my app interacts with dialers. Dialers call phone numbers until they get a connection with a person, then it drops to the agent (desktop) so the person can start talking. If your screen doesn't refresh instantly with the name of the person you are talking to, then the agent has to wait for the system to load. "Hi, I'm looking for... Uh... Hold on... System is slow today." Not to mention client pay/charge per minute alot of the time so the time adds up. The underlying technology of web development is counter intuitive for the app I'm making. So why would I want to embrace it?

So with that said. When I was reading the last webpage in my OP and seen that "web is the way to go", I disagreed. I think it has its point in the industry, but in my opinion it almost seems overused when a native app would be more appropriate. To me, it seems web apps are almost being overused these days for everything when it shouldn't be. Are people trying to push the square peg into the round hole and make it work for everything?

I've had people quit (I'm the manager) because I refused to convert everything over to Ruby with everything being a browser app because they read Ruby and web apps was the way to go on some blog. Oh, I can't forget Git as a source control method.

So now a few years later, I'm hearing it again in this forum and wondering if I'm just old fashioned like the guy who won't give up his record player and 8 tracks.

 

[sourceninja]

What I'm saying is as an application developer, I'm trying to find a good reason to learn web programming as well as I know application development. If I'm creating something that is external, to the consumer, such as viewing upcoming movie times at a theater, sure web page/web app all the way.

But typically in every place I've ever worked, they typically sell software to places such as brokerage firms, banks, remit/billing departments, etc. And web development seems like it would be a waste of time to explore. Unless you wish to expose your app to the outside world for the masses such as online banking.

I'm working on collection software at my place of employment, and I was told by the IT staff I need to make it web based (its the future!). But my app interacts with dialers. Dialers call phone numbers until they get a connection with a person, then it drops to the agent (desktop) so the person can start talking. If your screen doesn't refresh instantly with the name of the person you are talking to, then the agent has to wait for the system to load. "Hi, I'm looking for... Uh... Hold on... System is slow today." Not to mention client pay/charge per minute alot of the time so the time adds up. The underlying technology of web development is counter intuitive for the app I'm making. So why would I want to embrace it?

So with that said. When I was reading the last webpage in my OP and seen that "web is the way to go", I disagreed. I think it has its point in the industry, but in my opinion it almost seems overused when a native app would be more appropriate. To me, it seems web apps are almost being overused these days for everything when it shouldn't be. Are people trying to push the square peg into the round hole and make it work for everything?

I've had people quit (I'm the manager) because I refused to convert everything over to Ruby with everything being a browser app because they read Ruby and web apps was the way to go on some blog. Oh, I can't forget Git as a source control method.

So now a few years later, I'm hearing it again in this forum and wondering if I'm just old fashioned like the guy who won't give up his record player and 8 tracks.

Think on this. BYOD is becoming a reality. It sucks for IT people, but it just simply is happening naturally and outside of our control.

Do you want to write a win app, ipad app, osx app, android app? Or do you want to write a web app? Sure you say "We only support windows 7 and IE9", but then the president decides he wants a ipad and you have to draw straws to see who explains to the president why you need a whole project just to let him check his daily numbers.

If you are doing business logic, there isn't really all that much difference with the exception of a true "write once run anywhere" platform (especially that all browsers are basically feature compatible now).

I'm not saying it's the answer to all life's little problems. However, in IT it's always good to look at trends that are gaining traction and be aware and ready for them. It's much worse to be left out in the cold.

 

[Ken g6]

I see I forgot to mention something in my first post: separation. Web apps provide a good separation between logic (on the server, and sometimes with JavaScript), text and controls (HTML) and formatting (CSS, and sometimes JavaScript, but that is to be avoided). Once a web app is created, it's relatively easy to re-skin it with CSS and make it look completely different. You can even do this as an independent programmer. Likewise, you can change the text of your site (translation), and if your backend is well written, you might be able to change the HTML controls without changing the backend.

Now we need more to web apps than just the view/UI. What if you want to read/write to a database? Now you have to go through the web server. Using different frameworks here and there.

Well, the web browser does take care of the simplest case here. Every web browser can submit a form, and just about every backend framework has an easy way to handle form data.

I think the other attempt to respond to your "different frameworks" issue is called node.js. JavaScript here, there, and everywhere! I find it silly, but at least they're trying.

The only way to get decent response is if you poll frequently and with the web based app all those users/browsers polling takes a toll on the servers.

Technically, there is another way. The HTTP/1.1 spec allows a client to keep a channel open to the server. The client can keep a channel open and wait for any server responses that come back. Restarting the connection is necessary periodically, but it takes several minutes for a connection with no data coming back to time out. The downside to this, of course, is that the server has to keep a socket open to the client the whole time.

 

[beginner99]

What I'm saying is as an application developer, I'm trying to find a good reason to learn web programming as well as I know application development. If I'm creating something that is external, to the consumer, such as viewing upcoming movie times at a theater, sure web page/web app all the way.

I'm working on collection software at my place of employment, and I was told by the IT staff I need to make it web based (its the future!). But my app interacts with dialers. Dialers call phone numbers until they get a connection with a person, then it drops to the agent (desktop) so the person can start talking. If your screen doesn't refresh instantly with the name of the person you are talking to, then the agent has to wait for the system to load. "Hi, I'm looking for... Uh... Hold on... System is slow today." Not to mention client pay/charge per minute alot of the time so the time adds up. The underlying technology of web development is counter intuitive for the app I'm making. So why would I want to embrace it?

As mark already said. Web apps are good for some problems and desktop apps for others. In your case you need to convince your superiors that it is technical not feasible or impossible to do (I know that can be difficult if that guy believes everything he reads on some random blog or journal).

So if you only encounter problems that are not well suited for web apps it is understandable you don't like the "web app idea".

I've had people quit (I'm the manager) because I refused to convert everything over to Ruby with everything being a browser app because they read Ruby and web apps was the way to go on some blog. Oh, I can't forget Git as a source control method.

So now a few years later, I'm hearing it again in this forum and wondering if I'm just old fashioned like the guy who won't give up his record player and 8 tracks.

Well your probably lucky they left at least if you were in the right and they in the wrong. Can't tell that from here.

IMHO web apps (for internal use) are good for sharing existing data and adding new data. It is more consumer oriented, eg browsing and searching not creating. Like you would create a new 3D model in a desktop app and upload the result as VRML (or whatever) to web for other users to look at or search for.

Desktop Apps = "producing"
Web Apps = "consuming"

 

[mosco]

I am not even sure where to begin with this thread. I am not trying to be snobby, but if you have the means, I would quit your job, move to some place like San Francisco, Seattle, or Boston and trying working at a company that doesn't make software for insurance agencies and banks.

I too would quit a job if my manager didn't see the value of using Ruby or using GIT. And it would be an instant knock against any potential candidate for a job if they didn't have a Github/BitBucket login.

 

[brandonb]

I am not even sure where to begin with this thread. I am not trying to be snobby, but if you have the means, I would quit your job, move to some place like San Francisco, Seattle, or Boston and trying working at a company that doesn't make software for insurance agencies and banks.

I too would quit a job if my manager didn't see the value of using Ruby or using GIT. And it would be an instant knock against any potential candidate for a job if they didn't have a Github/BitBucket login.

Wow. I think I'll stick with what I'm doing. I'm sure its way more profitable and more career building than what you suggest. Thanks for your input though.

 

[mosco]

Wow. I think I'll stick with what I'm doing. I'm sure its way more profitable and more career building than what you suggest. Thanks for your input though.

Sorry to be blunt, but you asked if you were being old fashioned. There are reasons that the most innovate companies are switching to things like Ruby or Git that you brought up, or other technologies like Node, Backbone, Mongo, etc. And it's not because they are some kind of "fad". It is because they make sense.

I have never professionally developed in Ruby, but some of the smarted people I know wouldn't use anything different. My team uses exclusively for server side development. Lot of companies are using Ruby or moving to Ruby. Obviously it doesn't fit every use case, but it's gaining a lot of traction.

Things like Git are also where things are moving. I have used perforce and svn professionally, and nothing comes close to the power and convenience of git. It also is way easier to collaborate with people using git.

Don't get me wrong, I understand the importance of "applications". I am an iOS developer, so that's what I make. But there are so many reasons why building web apps make sense that you shouldn't ignore it.

 

[DannyBoy]

Just recently I was able to, for the first time, work with a top-shelf front-end designer and markup guy who produced bullet-proof html/css for a client site. All I had to do was turn it into cshtml views and the result was consistent across all browsers. It was awesome.

Yes we do exist, admittedly though good front end / UI engineers are few and far between.

 

[DannyBoy]

What I'm saying is as an application developer, I'm trying to find a good reason to learn web programming as well as I know application development. If I'm creating something that is external, to the consumer, such as viewing upcoming movie times at a theater, sure web page/web app all the way.

But typically in every place I've ever worked, they typically sell software to places such as brokerage firms, banks, remit/billing departments, etc. And web development seems like it would be a waste of time to explore. Unless you wish to expose your app to the outside world for the masses such as online banking.

I'm working on collection software at my place of employment, and I was told by the IT staff I need to make it web based (its the future!). But my app interacts with dialers. Dialers call phone numbers until they get a connection with a person, then it drops to the agent (desktop) so the person can start talking. If your screen doesn't refresh instantly with the name of the person you are talking to, then the agent has to wait for the system to load. "Hi, I'm looking for... Uh... Hold on... System is slow today." Not to mention client pay/charge per minute alot of the time so the time adds up. The underlying technology of web development is counter intuitive for the app I'm making. So why would I want to embrace it?

So with that said. When I was reading the last webpage in my OP and seen that "web is the way to go", I disagreed. I think it has its point in the industry, but in my opinion it almost seems overused when a native app would be more appropriate. To me, it seems web apps are almost being overused these days for everything when it shouldn't be. Are people trying to push the square peg into the round hole and make it work for everything?

I've had people quit (I'm the manager) because I refused to convert everything over to Ruby with everything being a browser app because they read Ruby and web apps was the way to go on some blog. Oh, I can't forget Git as a source control method.

So now a few years later, I'm hearing it again in this forum and wondering if I'm just old fashioned like the guy who won't give up his record player and 8 tracks.

For the record I've worked on 2 projects with full real time telephony integration (web browser based) built on a node.js stack - controlling all aspects of calls from within the browser.

 

[DannyBoy]

Sorry to be blunt, but you asked if you were being old fashioned. There are reasons that the most innovate companies are switching to things like Ruby or Git that you brought up, or other technologies like Node, Backbone, Mongo, etc. And it's not because they are some kind of "fad". It is because they make sense.

I have never professionally developed in Ruby, but some of the smarted people I know wouldn't use anything different. My team uses exclusively for server side development. Lot of companies are using Ruby or moving to Ruby. Obviously it doesn't fit every use case, but it's gaining a lot of traction.

Things like Git are also where things are moving. I have used perforce and svn professionally, and nothing comes close to the power and convenience of git. It also is way easier to collaborate with people using git.

Don't get me wrong, I understand the importance of "applications". I am an iOS developer, so that's what I make. But there are so many reasons why building web apps make sense that you shouldn't ignore it.

One word for git: cherry-picking.

Ok that's two(ish) words.