{"id":52640,"date":"2019-11-13T00:00:00","date_gmt":"2019-11-12T21:00:00","guid":{"rendered":"https:\/\/prohoster.info\/blog\/blog_prohoster\/xml-prakticheski-vsegda-primenyaetsya-ne-po-naznacheniyu"},"modified":"2020-02-18T14:00:25","modified_gmt":"2020-02-18T11:00:25","slug":"xml-prakticheski-vsegda-primenyaetsya-ne-po-naznacheniyu","status":"publish","type":"post","link":"https:\/\/prohoster.info\/en\/blog\/administrirovanie\/xml-prakticheski-vsegda-primenyaetsya-ne-po-naznacheniyu","title":{"rendered":"XML is almost always used for purposes it was not intended for","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"

\"XML
\nThe XML language was invented in 1996. Barely had it emerged when its potential applications were already being misunderstood, making it a poor choice for the purposes it was being adapted for.<\/p>\n

It would not be an exaggeration to say that the overwhelming majority of XML schemas I have encountered were an impractical or incorrect use of XML. Moreover, such applications of XML reflect a fundamental misunderstanding of what XML primarily is.<\/p>\n

XML is a markup language. It is not a data format<\/b>. In most XML schemas, this distinction was clearly overlooked, confusing XML with a data format, which ultimately meant that XML was an incorrect choice, as what was actually needed was a data format.
\n<\/a><\/noindex>
\nIf we do not delve into details, XML is best suited for annotating blocks of text with structure and metadata. If your main task is not working with a block of text, choosing XML is unlikely to be justified.<\/p>\n

From this perspective, there is a simple way to check how well an XML schema is constructed. Take, for example, a document in the assumed schema and remove all tags and attributes. If there is no meaning in what remains (or if an empty string is left), then either your schema is poorly constructed, or you simply should not have used XML.<\/p>\n

Next, I will provide several of the most common examples of poorly constructed schemas.<\/p>\n

<\/code><\/pre>\n
\nHere we see an unreasonable and strange (albeit quite common) attempt to express a simple \"key-value\" dictionary in XML language. If all the tags and attributes are removed, an empty string remains. Essentially, this document represents, however absurd it may sound, a semantic annotation of an empty string.<\/p>\n
<root name=\"John\" city=\"London\" \/><\/code><\/pre>\n
\nWhat's worse, we don't just have a semantic annotation of an empty string here as an extravagant way to express a dictionary \u2014 this time, the 'dictionary' is directly encoded in the root element's attributes. As a result, the specified set of attribute names on the element becomes indeterminate and dynamic. Furthermore, it is clear that all the author actually wanted to express was a simple 'key-value' syntax, but instead, they made the bizarre choice to use XML, forcing the use of a single empty element simply as a prefix for the attributes syntax. I come across such schemes very often.<\/p>\n
John\n  London<\/code><\/pre>\n
\nThis is already an improvement, but now the keys are somehow treated as metadata while the values are not. Quite a strange perspective on dictionaries. If you remove all tags and attributes, half the information will be lost.<\/p>\n
A proper representation of a dictionary in XML would look something like this:<\/p>\n
Name\n    John\n  \n  \n    City\n    London<\/code><\/pre>\n
\nBut if people have made the strange decision to use XML as a data format and then organize a dictionary with it, they must understand that what they are doing is inappropriate and inconvenient. Often, designers mistakenly choose XML for building their applications. But even more frequently, they exacerbate the situation with the senseless use of XML in one of the aforementioned forms, ignoring the fact that XML is simply not suitable for this.<\/p>\n
What's the worst XML scheme?<\/b> By the way, the prize for the worst XML scheme I've ever seen<\/i> goes to the configuration file format for automatic resource provisioning for Polycom VoIP phones. Such files require loading XML request files via TFTP, which... In short, here's an excerpt from one such file:<\/p>\n
<softkey\n        softkey.feature.directories=\"0\"\n        softkey.feature.buddies=\"0\"\n        softkey.feature.forward=\"0\"\n        softkey.feature.meetnow=\"0\"\n        softkey.feature.redial=\"1\"\n        softkey.feature.search=\"1\"\n\n        softkey.1.enable=\"1\"\n        softkey.1.use.idle=\"1\"\n        softkey.1.label=\"Foo\"\n        softkey.1.insert=\"1\"\n        softkey.1.action=\"...\"\n\n        softkey.2.enable=\"1\"\n        softkey.2.use.idle=\"1\"\n        softkey.2.label=\"Bar\"\n        softkey.2.insert=\"2\"\n        softkey.2.action=\"...\" \/><\/code><\/pre>\n
\nThis is not someone's failed joke. And this is not my invention:<\/p>\n