13.1 - Data on the Web

So welcome to Chapter 13. We’re going to talk about web services. And we’ve been talking about moving data using the request/response cycle and HTTP and urllib. Web services is really adding a layer of formalism on top of that. Where we’re just being a little more formal about how we do this and basically, at some point we’ll just switch from it’s moving data back and forth to these are are APIs, application program interfaces. And so, like we’ve said before, this request/response cycle that was originally for documents and images, has been used for data. And we have been coming up with ways to move data in a way that really have nothing to do with humans viewing them, but instead have programs producing and consuming this data. And the basic idea is you have data in a program, and so we’ve got two programs here, and they’re going to communicate across the Internet. So we might have a Python program that’s producing the data. Maybe it’s reading a database, maybe it’s reading a file. Who knows what it is. But inside it has a Python data structure, like a dictionary. And we want to send that across the network. Okay, and so the network is not Python. The network is not Java. The network is a data, I mean it’s data that’s going to go across. And so we have worked out, over the years, what we call the Wire Protocol, or how the data is put on the wire, or how the data leaves one system, transits a network, and then enters another system. And in that destination system, it’s not always Python. It could be another program. And so, perhaps, our Python dictionary in this other system needs to be a Java HashMap. And so, we can’t say that we’re going to send Python data across the network, and we can’t say that we’re going to send Java data across the network. We just have to send network on some format that we agree on. And so we have to argue about what the format is and say, okay we’re going to do this, and this XML, which is one of the wire formats. And it’s, okay we’re going to take this data that’s in a Python dictionary and we’re going to, XML looks kind of like HTML and it has less thans and greater thans as tags. And we’re going to send a person across the network, person that’s going to have a name and a phone number. That’s the data we’re going to send across. And we’re going to say, that’s our wire format. It’s not how Python thinks about it, it’s not how Java thinks about it. It is an agreed on intermediate protocol that is just text, right? It’s not internal memory. And the act of going from an internal representation on one computer out to a sort of interchange format is called serialization. And that has to do with the fact that, in the old days we had these wires, and we sent the data serially, across one character at a time. So it was taking, from the internal memory of the computer a format that we could sort of send one character at a time, character, character, character, character, character, so we called this a serialization format. And so, then the act of taking the data off of the wire and turning it into a new internal data structure, in the new environment, potentially in a very new language, is called de-serialization. So we take our internal structure, serialize it, send it across the network, then we receive it. We de-serialize it, and then we use it in this other programming language, in whatever structure makes sense, in that particular programming language. The two types of serialization formats that we’re going to talk about are XML and JSON. And so those are the two. And XML’s kind of like the older of the two, and JSON is the more modern of the two. XML is the more complex and, some would say, more rigorous of the two, and JSON is the lighter-weight version of it. So you take your Python dictionary, you produce JSON. You send JSON across the network as a string or a document, and then you receive the document, and then you turn it into whatever it is it’s going to be on that far side. So that’s the basic idea of agreeing on data formats. And so you argue here in the middle, we can argue, the Python people can come to the argument, the Java people, and the JavaScript people, they can all come and argue about what the best wire format is. And that kind of engineering of an interchange that is not particularly suited to any language better than any other language, is part of the argument of building these data oriented documents versus sort of human readable oriented documents. So, the first thing we’re going to talk about is XML. And, as the first of these two formats we’re going to talk about.