Related
I have a Coldfusion application, developed without any framework and almost no architecture.
I'd like slowly migrate some part of the application to some kind of java based web application framework.
The original application must still be the main application all the way until the end of the migration.
The application has user and sessions and a lot of functionalities, not easy to decouple.
I'm looking for different ideas.
For example I could try to develop a REST API backe nd and start to use it from the ColdFusion application until I'll have Coldfusion only as front end. This process must go hand by hand with database refactoring and migrations, ie. new API must use its own database, so I guess that decoupling of database will be necessary (and probably database synchronization issues will arise)
But I would like to switch also to a different front end technology.
The situation would be like: Users login in coldfusion, enter main dashboard in Coldfusion and next "some" functionalities will be handled by another framework. That means a template engine is able to understand the user, his roles, his permissions, reproduce the same graphical layout, but it should serve the content with another technology.
Final result must be that all functionalities are migrated to the new technology.
I mention Java as it is in some way related to ColdFusion, but any web application framework could be used in principle.
I also think that Coldfusion is used in the original application is not relevant. The fact that no framework is used, probably gives me more flexibility.
Any architectural suggestion is welcome.
I recently did this very thing. A spaghetti code app developed by a graphic designer with a CFML book, a MSSQL database so denormalized not even Bizarro could have wrestled it and won.
What we did was we left the old app in place, built a new one and tested it extensively, perfected it, made sure it was capable of evolving (and it has). Then we flipped a DNS switch and sent out an announcement that the new system is the only option going forward.
I then built an archive feature that provided read-only access to the mess that we left behind. Some day, if the app owners want to migrate that data, we can try that (it won't go well), but generally the vast improvements have convinced everyone that the old stuff is only needed in that read-only mode.
Obviously that means reporting only goes back to the first day of the new tool, but it was either that (with a 4 week dev time) or they get it all and it takes 8 months doing what you described.
It's worked very well for us, and it's probably been the best-received work I've done here.
I'd like to know if there's something like a dll or lib (with headers, and possibly documentation) that can be linked into a normal* win32 C++ solution so it can (by function call) notify Application Insights of page views and send metric values.
This seems like it should exist and be simple to find, but I'm just going round in circles on Google, finding documentation for reading values back out via a web API, or using C#, or projects that other people tried in the past but now have no source available.
This is for updating a diagnostic tool which I'm trying to avoid having to completely re-architect because the logging now needs to go to a slightly different service.
*ie. not Winforms, C++/CX, Xamarin or UWP. Straight, command-line c++.
Currently there is no Application Insights client library for native applications (c/c++). Though it's in backlog I don't think it will be prioritized higher than java/node.js. So, I would not rely on its release anytime soon.
Customers who wanted to send telemetry from native applications usually picked one of the following options:
Using REST API directly. If application is going to use only couple of telemetry documents then it should be straightforward. With the most complexity coming from own requirements such as persistent channel.
Host CLR + use .NET Application Insights SDK
Have .NET service running on a box and implement communication channel between native app and this service.
Depending on environment (for instance, #2 makes sense if app already hosts .NET; #3 makes sense if there is already installed .NET service in addition to native app) you can check whether on of these options will work for you.
Please excuse the noobiness of my question. I am mostly searching here for some directions and buzzwords to start digging from.
I spent some time developing an application in Python
Basically, it takes a bunch of images and creates a video out of it.
It i quite simple, and uses only a few libraries (opencv and nunmpy mostly).
I designed a small gui in gtk, but I think that it would be a good idea to offer the service over the web.
I think I could reuse some of my core and design a front end that people could access in their browser.
I only need a few data to get it running (images, an email)
The thing is my web dev skills are really close to 0, and I don't exactly know where to start from .
I don't plan on having hundreds of people a day on the platform.
People would connect, feed me with the data (link to a dropbox folder, google drive, whatever) and I would send them a message where it's finished.
If you could provide me with some names or links so that I could touch the field, I'd be really glad.
CGI is a fine option, but if you already have Python experience Django is definitely worth checking out (it falls in the category of rhooligan's #3 except it uses Python!). Django completely takes care of all of the database backend details for you, which is a benefit over simple CGI. It also provides easy-to-use pre-defined classes for handling file uploads, images, etc. It also has a great tutorial that will get you up and running. Just be careful about whether you're using version 1.3, 1.4, or the latest dev version, because some aspects of the framework have changed fairly quickly. Make sure that you're always looking at the right version of the docs.
Another handy service to keep in mind for doing something like image processing through a web app is a hosted cloud computing service provider like PiCloud. Unless you already have a private web server with lots of memory and processing power, these cloud services that charge by the ms are really cool. They also give you 1000s of cores which could allow you to do lot's of concurrent processing. They provide a nice Python API, and it has numpy and opencv pre-installed in both v2.6 and v2.7. (They use PyOpenCV, but you also have root access to install anything you want, so you can set up the "cv2" interface if that's what you're using--actually I just looked at your GitHub and it looks like you're using the old "cv" interface. You can also install any application you want on PiCloud--it doesn't have to be Python.)
You could start by looking into the Python CGI module and see if it will work for you. Then you'll need to do the following steps:
Decide on a webserver and install it, Apache is probably a good starting point.
Design the UI. Wireframe things out on paper paper. Figure out how you'd ideally want the users to go through your site and what you want on each page/view.
Your decision in #2 drives all the decisions from this point out. These days, most web applications are a combination of Web 1.0 and JSON/REST "services" (there's a couple of buzzwords for ya!). JQuery is a popular and widely used JavaScript library for developing the front end of your site. That would be another thing to look at. JQuery is completely independent from the back end and can be used with any type of back end (PHP, Ruby, Perl, .NET, etc)
Before you start flaming, I'm going to tell you that I am trying to convince myself that this is a bad idea.
Basically, I'm trying to create a website with some basic accounting functions. My friend, a consultant who only knows excel, asked if this could be an excel spreadsheet instead of a web interface.
I found myself thinking, why is excel not the better tool in this case? It has all the tools a web2.0 app has (scripting, access to a db, basic formatting), and is made for accounting. You could basically use excel to write a program that fetches data from an SQL database, populates some cells, and use formulas for the rest.
Is there precedent to using excel as one would use the browser to make an ajax-y web2 app?
Why is this a bad idea?
Update: just to clarify I meant this to be more a "rhetorical" discussion. I'm not sure why the rest of the message didn't clarify that...
Well, if you made it in excel, then it wouldn't really be a web application. It would be an Excel application.
Not that that's a bad thing though. If it makes more sense to build the app in Excel, then go for it.
The problems you'll have doing it in Excel will be the same as the problems you run into with any desktop vs. web application. For example, How do you handle deploying new versions of the app? How do you handle updating the app for new (or older) versions of Excel? etc.
ETA:
If you want to avoid the bugginess of Excel VBA and the headaches of managing a desktop application, you might want to look into the Google Spreadsheets API. You can use it to create/update Google docs spreadsheets on the fly, including formulas and lots of other spreadsheet goodness. Using a hybrid of web application and Google Spreadsheets might give you the best of both worlds, depending on what exactly you need to do.
After years of programming in Excel-VBA, the best answer that I can give you for not doing this:
Excel-VBA is buggy! It is probably the most bug-filled app that Microsoft has produced. It's great for some tasks, but forcing it to do a job for which it was never intended will lead to trouble.
I have a few spreadsheets that do similar things, (things Bill never intended), and without exception, they teeter on the brink of failure, and tend to crash with only the slightest provocation.
Sure, you can do it, but the headache is not worth it.
Excel was not meant to be used that way, so it will be painful.
Two better ideas for web-spreadsheet integration:
Take a look at Resolver One, a programmable Excel-compatible spreadsheet with an integrated web server.
Use Google Docs spreadsheets. There is an API to interact programmatically with them.
It's not a bad idea, but it does come with some limitations. If deployment isn't an issue for you, and you don't need "universal" access to the application, your solution will likely save you some time. A web application would certainly be a more elegant approach, particularly if you want to make the software available publicly.
The web is inherently based on HTML and added to with Databases, CSS and server-side languages and javascript (possibly others). As with most things, you should use the right tool for the job. If you want a website then you should use the tools for creating a website.
Excel is not meant to make websites, forcing it to do so will likely lead to more frustration than happiness.
Well, you're going to run it on the server side, you'd probably run into licencing and performance issues.
If you're delivering .xls to the client, then you'd need for all your clients to have Excel or something compatible. But you also lose control over your "site" - what's the point of visiting if you have everything you need locally?
So, the SQL statements would be on the client? That's rarely a good idea.
It's a bad idea for the same reason that writing a graphics editor with MS Access is bad, or coding a MMORPG using Powerpoint is bad :)
I would also say that once you stop using a browser as the client it ceases to be a "web application" - you are really just talking about an Excel sheet which fetches its data via HTTP.
+1 to everyone who said "an excel app is not a web app."
BUT... If you want to use the spreadsheet metaphor for server-side calculations for a web app, or if you want to access the library of financial functions that come with Excel from server-side code, you can use Excel Services. It's exposed via SOAP, interoperable, callable from any SOAP-capable platform.
It's server-capable. It does not actually load Excel on the server, but a non-GUI runtime of the functions.
I think that using Excel as Browser is bad idea, however i think better idea is to use Excel Control in Desktop App.
This way you can control the sql and connections. you can save data as often as you like. you could also implement some update mechanism. App would be more secure and harder to hack.
I think Excel freezes when you try to connect to some outside resource and this way you would control everything.
Did you start by getting user requirements for this application? It doesn't sound like they wanted a website to begin with. Sounds like they wanted an excel spreadsheet with macros.
Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
We don’t allow questions seeking recommendations for books, tools, software libraries, and more. You can edit the question so it can be answered with facts and citations.
Closed 4 years ago.
Improve this question
I'm building an app that authors would (hopefully) use to help them, uh.. author things.
Think of it like a wiki but just for one person, but cooler. I wish to make it as accessible as possible to my (potential) adoring masses, and so I'm thinking about making it a web-app.
It certainly doesn't have to be, there is no integration with other sites, no social features. It involve typing information into forms however, so for rapid construction the web would probably be the best.
However, I don't really want to host it myself. I couldn't afford it for one, but it's mostly that people who use this may not want their data stored elsewhere. This is private information about what they are writing and I wouldn't expect them to trust me with it, and so I'm thinking about making it a thick-client app.
And therein lies the problem, how to make a application that focuses mainly on form data entry available easily to potential users (yay web apps) but also offline so they know they are in full control of their data (yay thick-client apps).
I see the following solutions:
Build it as a thick-client Java app and run a cutdown version on the net as an applet that people can play with before downloading the full thing.
Build it as a Flex app for online and an Air app for offline (same source different build scripts basically).
Build it as a standard web-app (HTML, JS etc) but have a downloadable version that somehow runs the site totally on their computer. It wouldn't touch the net at all.
Ignoring 1 and 2 (I'm looking into them separately), I think 3 would involve:
Packaging up an install that contains a tiny webserver that has my code on it, ready to run.
Remapping the DB from something like mySQL to something like SQLite.
Creating some kind of convience app that ran the server and opened your browser to the right location, possibly using something like Prism to hide the whole broswer thing.
So, have you ever done something like this before?
If so, what problems did you encounter?
Finally, is there another solution I haven't thought of?'
(also, Joyent Slingshot was a suggestion on another question, but it's RoR (which I have no experience in) and I'm 99% sure it doesn't run under linux, so It's not right for me.)
I think you should look at tiddlywiki for inspiration.
It's a wiki written in JavaScript entirely self-contained in a single html file. You load it into your browser as a file:/// URL, so there is no need for a server.
I use it as a personal wiki to keep notes on various subjects.
Google Gears is used to offer a few of the google apps offline (Google Reader, Gmail, Docs and more).
What is Google Gears?
Gears is an open source browser extension that lets developers create
web applications that can run offline.
Gears provides three key features:
A local server, to cache and serve application resources (HTML,
JavaScript, images, etc.) without
needing to contact a server
A database, to store and access data from within the browser
A worker thread pool, to make web applications more
responsive
by performing expensive operations in
the background
Gears is currently an early-access developers' release. It is not yet intended for use by real users in production applications at this time.
If you're a developer interested in using Gears with your application, visit the Gears Developer Page.
If you wish to install Gears on your computer, visit the Gears Home Page. Please note, however, that Gears is not yet intended for general use.
But as you read it's still in early stages.
There is an additional option, and that is to use the new HTML5 offline application features, namely the Application Cache, Client-Side Databases, and Local Storage APIs.
Currently I believe that Safari is the only shipping browser to support any of these, and i believe it only supports the client side databases and local storage parts. The webkit nightlies support all of these features, the firefox nightlies support many of them (maybe all now?)
[Edit (olliej): Correction, Firefox 3 supports the Application cache, but alas not the client side DB]
We are using something similar to your third option to test our websites locally. Works just fine.
Our packaged webserver is not small enough to accomplish what you need, but then again we've not been trying to keep it small either. If you can package your webserver code into a small enough package I don't see why this approach would'nt work.
I think AIR is the way to go..
Have you checked into google gears?
Some pointers for solution 3:
for the GUI part, ExtJS seems really nice.
for the storage part, there is a nice javascript library that abstracts different storage backends: PersistJS.
Supported backends for PersistJS:
flash: Flash 8 persistent storage.
gears: Google Gears-based persistent storage.
localstorage: HTML5 draft storage.
whatwg_db: HTML5 draft database storage.
globalstorage: HTML5 draft storage (old spec).
ie: Internet Explorer userdata behaviors.
cookie: Cookie-based persistent storage.
Also, I think the moin moin wiki software has a desktop version that includes its own webserver. This stuff is easy in python, since batteries are included.
You might want to check out how they do it?
You could make a dedicated client using Webkit or Firefox's backbone. Some games use that solution for UI for example.
Or you could make a little webserver (I have a little webserver in Lua that I use for similar purposes, just a few megas with libaries and all). However if you take this route the biggest issue to consider is you don't want your webserver to depend on environmental variables, you want it to be totally autonomous. You should try to isolate all variables t o a config file and be done with it (bundle style)
Or you could use a Java client application to display the webpage
Or GoogleGears, but that's the same (almost) as Flex+Air. so choose Flex+Air if that's what you are familiar with
You didn't specify a language but I looked at Karigell a few years ago. It's Python web framework, similar to Django or TurboGears, but it doesn't have the overhead of those frameworks.
From my messing around with it, it seems like it would work for your purposes. It has a built-in web server (though you can use pretty much any server you want) and you can use any database that Python supports.
Plus, Python works well with Linux. :)
If you made the app a regular web app heavily reliant on client-side technologies (using DHTML and the likes of Google Gears to store data offline as already suggested) so once opened, there wasn't much interaction with the server, you could probably host the thing on a basic shared hosting account which wouldn't cost that much. That might be your easiest starting point as you wouldn't have to worry about all the issues with desktop apps such as compatibility with different operating systems, packaging up an install etc, yet you wouldn't need massive server resources behind it either.
You can use HTML, JS and whatever else in Adobe AIR and you'll have plenty of options of saving data locally, too.
in java world you could use jetty for a server, implement web app using your favorite framework and use hsqldb as a database - it lives entirely in your container (jetty). you can deploy preview app on the web and package downloadable offline version.
There's a portable distribution of Apache/MySQL/PHP (to place on USB keys):
http://portableapps.com/apps/development/xampp
This should be easily adapted to your needs.
You could also consider using XULRunner or Prism
They're the opensource technology that FireFox, Thunderbird and Joost are built on, and allows you to develop apps in XML and javascript essentially against the same rich api that FireFox itself has. And of course this is cross platform too, so it'd work on Mac/Linux/Windows...
Check here for more info:
https://developer.mozilla.org/en/XULRunner
I was thinking of doing something like this myself. My plan was to write app using django and write script that starts django's testing server and opens default browser on specified port. My plan was to use SQLite...
Also, it would be nice to pack it into one package, so users without django installed can run app without any dependecies...
My suggestion, as you pointed above, is to use a Wiki system to solve your problem. Now the question could be: Wich one?
You can use Trac, it is very simple and you can customize its GUI. But, if you prefer something more advanced please use MoinMoin. I used it for years, and IMO it is a very good and strong wiki system.
Depiste wich wiki you will choose, forget to write your web-app from scratch. According to yor question the best approach is to pick something that works and customize/modify it to fit your needs.