Philip Olson wrote:

On Sep 8, 2009, at 11:32 PM, Lars Torben Wilson wrote:

Lars Torben Wilson wrote:
Lars Torben Wilson wrote:
Daniel Convissor wrote:
Hi Lars:

In an earlier draft I had words to this effect, but it just got to be unwieldy, since it felt like if I explained that some people call class variables "properties", I should explain that others call them "members" or "member variables" or "attributes" or what-have-you. I'll work on it more. :)

When re-writing the Overloading section, it seemed "members" was the predominant word used already in the docs. It will be good to use a clear and consistent term througout the documentation. Considering the existing usage and the fact that "variables" already has a separate meaning for something else, I hope you'll consider using the word "members" here instead of "variables." Then explain that there are many other things people call them (and that they particularly hate being called "late for dinner").

Thanks,

--Dan

:)

Yes, this was along the lines of what I was planning to do. I haven't done a statistical analysis, but (as you noted) it seems to me that the preference already in the docs is to use 'member' as opposed to, say, 'property' or 'attribute'. I'll dig into some changes when I get home from work.


Cheers,

Torben

OK, I know I'm following up my own post, but I just wanted to put this out there: on the drive home, I realized what was bugging me about the use of the term "member" in this context: in typical OOP terminology, a method is also a member, as is a class constant. So while using the term "member" here might make things more consistent with some existing PHP documentation, my feeling is that the docs would overall be better served by fixing the existing usages to be more in line with common OOP terminology. Personally I like "property" (as opposed to, say, "attribute" or "field") but it's not a super-strong preference. However, while "variable" certainly isn't anything to carry forward, I don't feel comfortable using "member" either. Sorry about that--it just offends my sense of pedantry. ;) Anybody have any thoughts on this? I have no problem with updating the rest of the oop5 docs to match, no matter the outcome of our discussion.
Regards,
Torben

And, just to follow up to myself *again* (sigh, I know, I know), "property" would fit much better with the existing Zend engine internal naming structure, as well as match existing function names such as property_exists() and the names of various functions in the Object Aggregation extension.

The more I look into it, the more it seems like Overloading section (and any other part of the manual which refers to properties as "members") needs to be corrected, and this usage normalized throughout the documentation. In much of the documentation, in fact, it seems that the only things which would need to be corrected would be the links to the Overloading section and some more recent docs, as it looks like the oop5 extension may have been where this usage of "member" to mean "property" may have arisen. In many other parts of the documentation, "property" is the term of choice.

If we decide to indeed go this route (Philip, any thoughts on this?), I have already made the changes in most places in my checkout of the doc tree and can commit quite readily.

Oddly, I wasn't expecting this thread to go in this direction. :)

The above (properties) sounds like the proper route, and appears ready for commit.

Regards,
Philip

OK, sounds good. I've got one or two other outstanding commits to make (just committed a couple of them) before doing this cleanly, and now I must sleep, so I'll leave it until tomorrow to review my tree and leave a little time for anybody else to chime in who has an opinion.

Thanks Philip!


Torben

Reply via email to