Talk:Mm'-type filter

Citations

I converted the (3) citations to template form but this was immediately reverted by User:Spinningspark. While I am aware that discretion is used when choosing templates over not, my reasons are: This allows greater semantic content. The parts of the reference are labeled, links are provided to identifier numbers (OCLC/ISBN etc). Best of all, the code doesn't appear as a huge mudge inside a paragraph in the code editor, it is outsourced to the references section, clean and easily edited, with each parameter named and on its own line. It also frees the citation from the constraints of the author's style - in future, there can be options to view references differently (there are hundreds of reference styling formats: maybe I want APA style or maybe ISO style), convert to Bibtex references, or other such features. This is only possible if the information has sufficient semantic content, which the template provides.

As for the patent ref, there was a mistake causing it to not show the inventor's name, one line should read:

| inventor=Zobel, O. J.

rather than

| invent1=Zobel, O. J.

Other than, that, it provides a link to the esp@cenet database, where other versions of this patent, citing patents and a copy of the patent can be found. The advantage of a template is that it can be changed if the espacenet access method changed (e.g.the URL changes) or a different database is used in future. With a direct link, the link will just break.

I also added information (OCLC numbers, ISSN numbers, links to the journal pages, etc), which were not present before.

So pros:

• Clean markup, easy to edit for everyone (not just the original author).
• Non-inline markup, doesn't put huge blocks in the middle of paragraphs.
• Adds semantic information.
• Places links to terms like ISSN and OCLC, as well as allowing the reader to go and look the identifier up in a database.
• Provides links to the information that can be easily changed centrally should a database change.
• With added semantic content, the style can be flexible, and dependent on user preferences in future.
• References can be standardised between articles. You'd find the same in any other encyclopedia, journal or book, not a mish-mash of different author's styles.

Cons:

• The author doesn't have exclusive control over the exact, precise appearance of each letter, digit and comma of each citation (but if he really has a problem with it, he can go and ask at the template page).

I don't think there is much more to say, except that I performed this change along with many other improvements (such as nice graphs to replace the horrible Excel graphs that were here before, see also templates, etc) that were also reverted. As for the claim "I object to littering the article with templates", I think Spinningspark has horribly missed the point of a template: they allow information to be provided so that it can be kept coherent and consistent. They are not a bad thing in themselves, they are a powerful tool for adding meaning to the code and allowing restructuring of the information. (for the var tags, see below, I was just trying them out). Inductiveload (talk) 15:33, 26 November 2009 (UTC)

It is not (or should not be) one of the aims of an article to add semantic content to the code. Rather, we should be endeavouring to add semantic content to the article, the thing our readers can actually see. The code in an article is not a plaything for programmers, it is to aid editors in the task of constructing the article. If an editor has already succeeded in writing the article without the need to resort to templates, it is counterproductive to retrospectively insert them. This is my view for all kinds of templates, not just cites, although I do not deny that templates can be useful to an editor who does not already know the proper way of formatting citations. I don't think you understand how intimidating a page full of code can be for a new editor. I deal with a lot of new editors both on and off line and know this to be true - that a lot of code in an article will discourage a newbie from editing even when they definitely have something worthwile to add.

Templates can be remarkably frustrating things, refusing to produce the result you know should be there, despite multiple attempts. Yes, I know I could read the template documention, yes, I could go and ask and in any case I am not a complete incompetent in template code, I could read the code and figure it out myself. But there is no point if I can just write what I need to write in plain text in a much shorter time.

Is this idea of cites displayed according to user preference actually happening? Can you point me to a discussion or consensus, or is it, as I suspect, just a pipe dream. My view is that the only reason you would need a cite in different styles is if you were downloading it to use elsewhere. In which case the cite style should be selected at the point of download, the view in the article should be an editorial decision like everything else on the page. Please remember that the primary purpose of citations is for verifiability of the article, not as some sort of academic index.

Can I offer you a compromise? Author,(date),page style refs in the body of the article (plain text, no templates) between ref tags, and fully templated refs in a separate "Bibliography" section. As I said on your talkpage, I think the new graphs are a great benefit to the article and should be retained, it was just that a partial revert was not possible without a deal of effort and it was wiser to reach consensus first. SpinningSpark 17:04, 26 November 2009 (UTC)

OK, that compromise is acceptable. I just found out about the Harvard citation templates, and I just thought they were a neat way to add an automatic link from the footnote to the citation in question: without too much extra markup, you can just click and it highlights it for you, rather than looking through the list of references. But I don't particularly mind if you ditch these, as there are not too many citations to look through anyway.

Do you want me to do it, or would you rather put the bibliography section in order? Inductiveload (talk) 17:50, 26 November 2009 (UTC)

I'm fine with you doing it. SpinningSpark 19:20, 26 November 2009 (UTC)

Var tags

OK, this one I admit was a bit iffy, but var tags do also add semantic content to the code - it explicitly marks out an item as a variable, allowing the browser or renderer to display accordingly. I was just trying it out, because I thought it would be a neat way to show what is a variable and what is merely italicised. I'm sorry if I fell foul of Article 3.132.12a of the Spinningspark Manual of Style, but since it isn't made public it's a bit hard to keep in line. Inductiveload (talk) 15:33, 26 November 2009 (UTC)