AW: AW: AW: [PEAR-DEV] [RfC] XML_Serializer

Hi, > Your description above "creating XML files from PHP data structures" > implies that Josh's suggestion on a name is better than > XML_Serializer. I still insist that this class does neither generate data nor manage it. All it currently does is serialize data to XML any a customizable way. Basically you could say that it can be used for something similar to WDDX, but it can also be used to create RDF files, DocBook files or XSLT files. All it does is to take a data structure the user is familiar with (arrays or objects) and serialize it to an XML document like you would expect it to be serialized. Types are ignored unless you enable type hints, which is something that's only needed if the Unserializer is usedto read the file. What you get is something like a print_r with XML tags :-) > If you use "serialize" in the name, all users will > associate it with > the serialization that they know, which is serialize and > wddx_serialize_*. If you don't want this association, which > you clearly > don't by the capital "NOT" :), then the name makes no sense. Of course, it can be used for this kind of serialization, but that's not the sole meaning of serialization. When transforming an XML document with XSLT and then sending it to the browser the XML document is transformed in memory and then _serialized_ to a string. This has nothing to do with WDDX but still _is_ a serialization! > You should > call it XML_Data_Generator as Josh suggested, or > XML_Data_Manager, since > you are planning to add an unserialize. If you have the > subclasses that > you mention, people will understand what they do in the context of > managing data. To imply that XML is not about data > interchange doesn't make sense, that is its sole purpose. It's not about the data interchange in the sense of WDDX, where you want to preserve data structures and allow different programming languages to read the data. > I am still -1 without documentation, but will +1 when I see > the solution > I describe below As I said before, I'm willing to write documentation and include examples for it, if you take a look at our site (http://www.php-tools.net) you'll clearly see, that we do not publish classes and let the users worry how and where to use them. > -the problem that this package solves (the reason this is a problem > should be illustrated with at least 1 example) > -the way this package solves that problem > -the differences from other similar solutions (explicitly mention > wddx_serialize and serialize) > If half of the developers on pear.dev don't understand the purpose of > your package on the initial proposal, that should tell you something > about the clarity of the purpose, not about the intelligence of the > pear.dev developers. We can't all be idiots, even if I > sometimes am ;). I think there's a problem in what some of the users expected from the package, but this cannot be only because of the name. As I will include an unserializer in a future version, it will do, what everybody expects: it can either be used to create various XML formats as well as create XML files that can be read be the unserializer and be restored as the initial data structure. Stephan