|
A kind of...I have to see in more details what we talking about...I may be able or unable to help...I can't know, but I always ready to help...
I'm not questioning your powers of observation; I'm merely remarking upon the paradox of asking a masked man who he is. (V)
|
|
|
|
|
Kornfeld Eliyahu Peter wrote: I have to see in more details what we talking about
Well, this is certainly an "on-the-fly" design, but I intend to put some initial planning documentation together. I'll send you a direct email (I have your email address) with a writeup in a day or two. I'd like to have some sort of bare-bones scaffolding in place by middle of next week, which should be interesting for me to ramp up on ASP.NET/Razor in the most efficient way possible.
Marc
|
|
|
|
|
Hi Marc,
I've been doing web stuff for several years now, having moved from desktop dev (MFC, WinForms, WPF, Silverlight) to ASP.NET web stuff.
Here's my preferred web stack:
- Bootstrap. CSS framework. It's a consistent style for your whole site. Makes it easy to build responsive sites that work well across all screen sizes. It makes it easy to do things like modal dialog boxes, gives you a consistent appearance across your whole site, and gives you some nice common web components such as drop-downs and navbars[^].
- jQuery - DOM manipulation framework, helps you do HTML manipulation consistently across all browsers. It's required by Bootstrap and still quite handy when you need to manipulate HTML elements by hand.
$("#someHtmlElement").text("some new text here!"); - KnockoutJS - Databinding and MVVM pattern. It lets you do things like this:
var myViewModel = {
foo: 42
};
ko.applyBindings(myViewModel);
<label data-bind="text: foo"></label> - ASP.NET MVC and ASP.NET WebAPI - MVC (with Razor) for server-rendering your HTML, and WebAPI for fetching data asynchronously.
I personally stay away from things like UI control frameworks. I find them to be vestiges of the desktop world that really aren't necessary in the web stack. There are some decent ones out there, such as Telerik's Kendo UI[^], but it's not really necessary.
One more thing. The hot new sexy thing everyone's worked up about is Google's AngularJS[^]. I've used this now on my last 2 web projects. Like Knockout, it provides data binding. Additionally, it's a full fledged MVC framework in JavaScript, providing client-side routing, data binding, seperation of concerns between view (HTML), presentation logic (controllers in JavaScript), and data services (data fetching AJAX calls). Use it if you're building a dynamic web app; e.g. where the UI changes often and shows live data without having to do server postback/page refreshes.
modified 22-Oct-14 14:45pm.
|
|
|
|
|
Judah Himango wrote: There are some decent ones out there, such as Telerik's Kendo UI[^], but it's not really necessary.
Yes, I've been watching what's going on there.
Judah Himango wrote: The hot new sexy thing everyone's worked up about is Google's AngularJS[^]. I've used this now on my last 2 web projects.
Yup, I'm looking at that. What ever happened to node.js? I see no one has been suggesting that?
Marc
|
|
|
|
|
NodeJS is JavaScript on the server, e.g. a replacement for things like ASP.NET, Ruby on Rails, etc. I personally don't want JavaScript on the server.
In fact, if I can tell you a dirty little secret, when on the client, I prefer TypeScript (or even CoffeeScript[^]) over raw JS. JS is missing a lot of things to make it a nice language for large apps. Languages like TypeScript and CoffeeScript help it in that regard while still compiling to plain ol' JS that runs in everybody's browser.
|
|
|
|
|
Judah Himango wrote: NodeJS is JavaScript on the server, e.g. a replacement for things like ASP.NET
Ah! That helps.
Judah Himango wrote: personally don't want JavaScript on the server.
Agreed.
Judah Himango wrote: I prefer TypeScript (or even CoffeeScript[^]) over raw JS.
I will check those out.
Thank you!
Marc
|
|
|
|
|
Cool. I know you're a C# guy, so TypeScript will make you feel right at home. Try it out in your browser[^].
TypeScript is created by Anders Heijlsberg, the creator of C#. Built into Visual Studio these days, so you can just Add New Item->TypeScript file, then start writing things like:
class MyAwesomeClass {
constructor(firstArg: string, someOtherArg: number) {
}
someFunc(answer: number): string {
return "The answer is: " + answer;
}
}
|
|
|
|
|
Marc Clifton wrote: What ever happened to node.js
I had the feeling you are not a fan of JavaScript, so why use JavaScript on the server too...
node.js is for writing server side code (like ASP.NET code behind) in JavaScript...It has good features, but it is not for everyone and for sure not for a site you talked about...
I'm not questioning your powers of observation; I'm merely remarking upon the paradox of asking a masked man who he is. (V)
|
|
|
|
|
 for jQuery and KnockoutJS
Jeremy Falcon
|
|
|
|
|
Judah Himango wrote: The hot new sexy thing everyone's worked up about is Google's AngularJS[^]
This[^] (knockout vs angularjs) is an amusing read, especially the pictures at the end of the article! Thought you might enjoy it.
Marc
|
|
|
|
|
Heheh. Their analogy via pictures is funny and accurate: Knockout is a tiny paddleboard (it really just does data binding!), whereas AngularJS is the big yacht that does data binding, routing, MVC, separation of concerns, dependency injection, and much more.
Great analogy.
|
|
|
|
|
If you're going to build a SPA (which it doesn't sound like he is), you should seriously consider AngularJS. "Those who don't know angular are doomed to repeat it." 
|
|
|
|
|
I dislike MVC, because it is unnecessarily complicated, and will take (me) longer to develop and maintain.
But if you start today, MVC with Razor is the way to go. Everything else is just outdated. MVC razor is the most recent standard for working with asp.net. (at least it was two months ago, gosh things change rapidly)
Leo
|
|
|
|
|
Leo Muller Rap wrote: it is unnecessarily complicated, and will take (me) longer to develop and maintain
Uhhhmm. I think this is just your opinion. I've worked many years with web forms, before moving to MVC.
MVC is not complicated, once you grasp the basic principles it is much easier to work with. No funny business trying to figure out how to make a web control generate an HTML that works. It's much less error prone.
It provides separation of concerns by separating the view from everything else. This makes it much easier to maintain as we can clearly separate logic from user interface. Which leads to a few more advantages:
- Unit test your user interface
- You can have a team specialized in UX working in the project and they do not need to know anything else but HTML/CSS, etc to contribute, while you work in the logic at the same time.
- You can control much better the HTML output to deal with some specific scenarios.
After my intial learning curve, all development I've done in ASP.NET MVC has been much faster than WebForms. I also spend time a lot less with bugs that are hard to work around in Web Forms.
Since the user interface is not bound to the framework, it integrates much better with the myriad of tools and frameworks available on the web.
And finally, thank the Lord Almighty, we don't have to deal with update panels anymore.
To alcohol! The cause of, and solution to, all of life's problems - Homer Simpson
----
Our heads are round so our thoughts can change direction - Francis Picabia
|
|
|
|
|
Hi Marc,
as it has been mentioned here most probably ASP.NET MVC seems a way to go.
The ASP.NET webpages is slightly lighter and not that structured as MVC (which could be good if you want to ship something quickly that is pretty simple).
ASP.NET webforms are still there but it is MVC that is catching a big momentum (if talking about Microsoft). Other non MS frameworks I heard of (but never touch them), your lovely ruby on rails and maybe java spring...
Some people here are mentioning WebAPIs with rich client side javascript frameworks (AngularJS, typescript, coffeescript, etc.). I would use these if planning to do some apps on mobile devices. I would not bother if targeting purely website interface. Well, I think javascript is unavoidable today, but WebAPIs are not always needed I think. But then again, if you want to play with the techno, this can be good opportunity to lay the hands on
Last thing - the database. ASP.NET MVC seems to be working fine with multiple db systems (MySql, MongoDB) but if you are Microsoft guy, I think there is no reason to go further then one of MS SQL mutations.
However if you are planning to use some ORM frameworks (Entity Framework, NHibernate, etc.) one thing to decide first would be the famous discussion Code First - Model First - Database First approach. Recently it seems that trend is Code First approach. The other two seem to me slightly less popular on discussion boards nowadays.
Just my 20p to discussion 
|
|
|
|
|
Bajaja wrote: I would use these if planning to do some apps on mobile devices.
That is the intention.
Bajaja wrote: one thing to decide first would be the famous discussion Code First - Model First - Database First approach.
For the back end, I prefer my own database solutions, but yes, there's no reason not to go with SQL Server, as it's provided automatically by any number of ASP.NET hosting sites.
Personally, I don't buy into any one of those paradigms as a fixed law of coding, and I rarely separate "model" and "database" - to me, they are both the same thing. So, depending on what I need to prototype, I might work on the UI first or the model first or do some general code stuff.
In this particular case, spending some time thinking about the UI and the model is what seems most reasonable first, as I want the site to be able to handle DPW's from more than just one state and municipality. I also want to consider how that affects users per DPW, administration, state-wide admin, and possibly country-wide admin, etc. Getting that stuff "right" from the get-go will mean a lot less BS "refactoring" later on.
Bajaja wrote: Just my 20p to discussion
Well worth it! Thank you for the input.
Marc
|
|
|
|
|
I want to add a few things to enforce what a few already have said and add my little two cents:
UI:
1 - Yes, definitely MVC/Razor. You will learn the delight it is to work with when you get there. Separation of concerns, unit testable UI, complete control on the HTML generated.
2 - jQuery is a must, typescript is great also to do Javascript with OOP (including generics). AngularJS will be very helpful, but I think you can leave that for a later project if you feel overwhelmed.
Database:
1 - I would definitely go for entity framework. It's very mature and if used carefully will save you a lot of development time and will not be a problem performance wise. Do go for code first. Of course you should experiment with the other ways of doing EF, but I would prefer you were spared of the problems of going the other ways. I also recommend using migrations with EF CodeFirst.
Services:
1 - WCF is the way to go, but you probably know that already.
I think these are the basics to get started. Enjoy.
To alcohol! The cause of, and solution to, all of life's problems - Homer Simpson
----
Our heads are round so our thoughts can change direction - Francis Picabia
|
|
|
|
|
Fabio Franco wrote: You will learn the delight it is to work with when you get there.
I hope so. I've done a lot of RoR so a lot of Razor looks familiar.
Fabio Franco wrote: jQuery is a must,
Indeed. I've played with it a little, again in the RoR environment.
Fabio Franco wrote: AngularJS will be very helpful, but I think you can leave that for a later project if you feel overwhelmed.
That's the conclusion I came to last night. I'm having a hard enough time finding a decent tutorial on getting Bootstrap folded in. Found this[^], but I may be making this too complicated -- in RoR, it's a simple matter of dropping in the css and js files into the appropriate folders.
Fabio Franco wrote: I would definitely go for entity framework.
Well, in this department, I have my own DB frameworks that I prefer to use.
Fabio Franco wrote: Do go for code first.
Hmm, I have some moderately complex modeling issues to consider, such as supporting DPW's in different municipalities and states. I prefer to have a solid model in place, perhaps even with some "scenario tests" to vet out the design, before doing too much coding / UI development. However, when the model is obvious, yeah, I often end up prototyping the UI first.
Thanks for the great feedback!
Marc
|
|
|
|
|
Marc Clifton wrote: Found this[^], but I may be making this too complicated
We may be talking about different things. AngularJS and knockout is about databinding MVVM way, that can improve productivity and maintainability when working with data models on the client side. The twitter bootstrap works like that in with .net mvc, it is as easy as dropping the js and css. If I am not mistaken it is even part of MVC project template already. If not, you can set it up easily using the nuget package manager. If you don't know about nuget package manager, please look into it, it is a life saver.
Marc Clifton wrote: Fabio Franco wrote: Do go for code first.
Hmm, I have some moderately complex modeling issues to consider, such as supporting DPW's in different municipalities and states. I prefer to have a solid model in place, perhaps even with some "scenario tests" to vet out the design, before doing too much coding / UI development. However, when the model is obvious, yeah, I often end up prototyping the UI first.
I think we may be talking about different things again. Code First, is and Entity Framework style of mapping Object Entities to database tables. It's like having a object oriented database. You can define complex models and relationships (including inheritance) with EF Code First that will later translate to a database. It can represent your domain model, although it's not always recommended, depending on your architecture. UI Model prototyping is something else, this is MVC's M stands for. It's something else. EF Code First entities are another layer, either on the data layer, or domain layer.
Marc Clifton wrote: Thanks for the great feedback!
Glad to help 
To alcohol! The cause of, and solution to, all of life's problems - Homer Simpson
----
Our heads are round so our thoughts can change direction - Francis Picabia
|
|
|
|
|
As others are saying, MVC/Razor, jQuery, Bootstrap.
I don't have experience with Telerik's controls in the MVC world, but I can say their .NET controls are execellent, with great support both from the user community and Telerik themselves.
|
|
|
|
|
Microsoft stack is: ASP.NET (with or without MVC), Razor, IIS and Entity Framework/MSSQL. But what about vNext? Seems to me that vNext is Microsoft's answer to Nodejs and looks to be a future 'upgrade' for ASP.NET (BTW Nodejs isn't written in Javascript, it's written in C and just translates JS). IIS is a slippery pig as far as I'm concerned but I still use it. The Javascript stack is: Angularjs (which is amazing), Nodejs (usually with Expressjs on top) and either SQL or Mongodb for the db. Knockout and especially Breezejs seems to be the ORM of choice. As far as open source goes, the creme de la creme rises to the top and isn't as 'open' from a development perspective as you might think (being controlled by the original authors which is now often Google, Amazon, Sencha, MS and a small group of super guru's). Open source it not a free for all. Typescript is a nice enhancement to JS but I don't like working with anything that's 1.0 (for production). I've been developing in C#/.Net since 2001 and now currently learning the Javascript stack and it's been PAINFUL! But, Javascript isn't a toy language anymore and is the language of web development going forward, just ask Microsoft! Cheers to all.
|
|
|
|
|
I'd start with "designing" (wire framing) the thing first before deciding on a "technology stack".
Didn't see any thoughts on "hosting" either. It may be "pro bono" but at some point there's going to be a cost of ownership.
Maybe there's a "SharePoint Online" in your future.
|
|
|
|
|
Gerry Schmitz wrote: I'd start with "designing" (wire framing) the thing first before deciding on a "technology stack".
Good point, and yes, that's my plan.
Gerry Schmitz wrote: Didn't see any thoughts on "hosting" either. It may be "pro bono" but at some point there's going to be a cost of ownership.
True, but from my prior research, it looks like I can find hosting for around $10/mo, which is pretty darn cheap, and given that the initial # of users is going to be about 10, I'm not too worried about performance at the moment.
Marc
|
|
|
|
|
It wasn't so much the "$10/mo" that I was thinking of (there are even cheaper options around).
It is the support and maintenance that comes afterwards.
While you may become a master of all the components you plan on incorporating, the fellow(s) that come after you may not be.
It appears that you are not being asked to follow any type of corporate "standard" (assuming there is one). Corporate standards usually imply some sort of internal skill sets that can later be called upon, if necessary.
An application that is not or cannot be supported, maintained and updated when necessary will rapidly fall into disuse.
(I like / hope to think the stuff I develop winds up having a long life).
That said, it's wonderful to be in a position where one has control over most, if not all, aspects of a project.
|
|
|
|
|
Gerry Schmitz wrote: It is the support and maintenance that comes afterwards.
Indeed, and what I'm hoping will result from what is essentially a prototype (or, perhaps a better word is "pilot program") is the value will be seen and there will be actual funds allocated for support and maintenance.
And yes, I hear you on the standards, etc.
So, it'll be interesting to see how this evolves over the next 6 months or so!
Marc
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
