Web API is a feature of the ASP.NET framework for simplifying the process of building RESTful HTTP services. The tool’s portability allows it to be used in services that are intended to serve across platforms, device and browsers, and users can create endpoints that are accessible via descriptive URLs and HTTP verbs.
In this tutorial on DeveloperHandbook.com, the author shows followers how to create a RESTful Book Store API that authenticates and authorizes requests, exposes OAuth2 endpoints, and returns data about books and reviews.
The set up occurs in Visual Studio where we open the packages that should have been added to Web API already, with the Entity Framework for data persistence. Followers can add CORS support if they wish, then create the data structure to represent books and reviews. The API Controller is set up to use the Entity Framework’s Seed method to expose data, with a RESTful endpoint that retrieves all of the book data.
The author then handles authentication and authorization using OAuth and JSON Web Tokens (JWT), discussing the enabling of OAuth and formatting of JWT in detail before getting into testing. All code is provided so followers can build the same RESTful API, with instructions to copy the sample code exactly to avoid unpredictable behaviour.
Well, “most popular” is a risky claim. Especially when it comes to programming languages: there are ardent admirers for each language and each language has its own niche where it rules.
The order of popularity changes based on what data you look at.
Before we get into the details of how we at Chargebee compiled the rankings, let’s look at some of the ways in which you could arrive at the list of top programming languages.
You could do a broadly based rating using multiple parameters such as search engine results like Tiobe did, or if you are popular among the devs like the folks at StackOverflow, you could decide to get it right out of the horse’s mouth.
You could also look at it based on job postings, which is what Coding Dojo did.
Or, you could have a different take on it — you could look at the number of API libraries that have been released for that language, like they do at ProgrammableWeb.
Collating them brings up a pattern, the odd one being ProgrammableWeb’s list:
| Popularity Ranking | Tiobe | Stack Overflow | Coding Dojo | ProgrammableWeb |
|---|---|---|---|---|
| 1st | Java | JavaScript | Java | PHP |
| 2nd | C | Java | JavaScript | Python |
| 3rd | C++ | PHP | C# | Ruby |
| 4th | C# | C# | Python | C# |
| 5th | Python | Python | C++ | Java |
Sources:
We decided to add one more angle to it by looking at the actual API requests from the apps that have integrated with Chargebee using the language-specific SDKs.
Even though you could use Chargebee as a standalone service for your billing needs, most SaaS companies integrate with us tightly to leverage our subscription billing features, and they tend to use the SDK of the language that their service is built with. So when we exposed our REST flavored API, we also released SDKs for the “popular” languages.
Our judgement of the “most popular” is based on the actual number of merchants using a particular language library to integrate with Chargebee for their billing.
We specifically took into account only the merchants who are actively making API calls to the billing system. We have excluded requests that come to the sandbox accounts.
Just like browsers add information about themselves (including its version) in the “user-agent” header, each of Chargebee’s client library adds its information in that header in http requests sent to the API endpoint, which in turn is recorded as a part of the access logs in the server.
A sample user-agent header sent by php client looks like this:
Chargebee-PHP-Client v2.0.2
For each request, we can identify the SDK used along with its version. We use Splunk to analyze our logs. A simple query gave us the results.
Before we jump into the data, let’s first look at the languages for which we have released the official SDKs and our take on their popularity.
The Web runs on PHP. Well, at least most of the popular content management systems like WordPress are based on PHP. So it might seem like the obvious choice at first. But giving it deeper thought, we expected the API users to be the “hip” startup crowd. We expected them to use more of the newer and cooler languages like Ruby or Python, rather than PHP.
That Ruby On Rails is popular among startups is no secret. We expected Ruby to come closer to PHP in terms of usage.
Python is another language favored by startups. We expected it to be slightly less popular than Ruby, but a lot more popular than Java.
We released the Node JS library based on few customer requests. It was the last language flavor we released. We didn’t expect much usage as the server side js platforms are still new in comparison with other platforms.
It has fallen out of favor with the startups. It is now considered enterprise-y. Too verbose and bloated. We expected very low usage.
Most of us at Chargebee come from the Linux background right from the beginning. And most of us have switched to Mac now. There are few Linux holdouts (like me). But MS and its developer ecosystem have always been an enigma to us. We weren’t sure of the number of users would use the .NET library when we released it.
We expected PHP to lead, closely followed by Ruby and Python, with Java trailing as the distant fourth.
A year ago, we looked at a slice of data spanning three months. And the usage pattern is as below:
As we can see, PHP still rules the web, followed at a distance by Ruby. The surprise is that the usage rate of .NET is more than that of Python.
We would have released the .NET library before the Python one, had we had this insight earlier.
No major changes. But PHP seems to be losing some points to Python and Ruby. NodeJS has increased and Java has dropped.
If we look at our findings, we could see that it is more in alignment with ProgrammableWeb’s rankings.
The stark point is the popularity of PHP when compared to other languages.
We understand that our usage based dataset might be biased due to the profile of the customers using Chargebee’s billing solution.
But if behemoths like Paypal and Stripe publish their findings, it would provide a more solid dataset as most SaaS businesses would at least have a payment gateway integrated, if not a full-fledged billing solution like Chargebee.
Do you have a dataset that’s different from what we’ve shared? We’d love to know!
This article was republished with permission from the author as part of our contributed content program. The original article can be read here.
