In preparation for "clean" URL rewrites

[Gigablah]

Now that we have an increasing number of sites linking to our content pages, there's a pressing need to finalize our URL format. My plan is to turn an address like this:

http://www.vgmdb.net/db/albums.php?id=14

Into this:
http://www.vgmdb.net/db/albums/id/14-einhander-original-soundtrack/

For listings:
http://www.vgmdb.net/db/albums/title/0-9/1/
http://www.vgmdb.net/db/albums/title/a/3/
http://www.vgmdb.net/db/albums/catalog/s/5/
...

And similarly for artist pages:

http://www.vgmdb.net/db/artists/id/294-kenichiro-fukui/


And so on. I don't plan on changing the URLs for the forums or submission pages -- only for the pages that display primary information.

There'll be 301 redirection for legacy and malformed URLs -- for example, accessing

http://www.vgmdb.net/db/albums.php?id=14

or
http://www.vgmdb.net/db/albums/id/14-gibberish-here/

will get you redirected to the proper address. This will prevent URLs from being broken whenever the album title (and the corresponding URL portion, or "slug") gets changed.


Any comments, particularly regarding the proposed URL format?

[Secret Squirrel]

Will there be any problems with having some of the more unusual symbols in the titles, or will those all be omitted from the URL? I can't find any examples, but I know there are some with stars or hearts.

[Gigablah]

The URI syntax restricts the characters that can be used. All accented characters will be converted to non-accented ones, "&" will be converted to "and", and punctuation and other symbols will be removed. Spaces will be replaced by dashes.

Some examples of the conversion:

"Donkey Kong Country 3 ~ Dixie's Double Trouble! Original Soundtrack (US)" => "donkey-kong-country-3-dixies-double-trouble-original-soundtrack-us"

"[-Aka-] Complete Edition Original Soundtrack - Primitive" => "aka-complete-edition-original-soundtrack-primitive"

[Kaleb.G]

Looks good to me. Reminiscent of the way GFF is handled.

[Secret Squirrel]

I don't know much about SEO, but I'd say that this scheme looks like a fine one.

[Gigablah]

I've done a bit of re-thinking about this.

I'm now considering scrapping the title slug completely, and just having ID numbers like
http://www.vgmdb.net/db/album/14/
http://www.vgmdb.net/db/album/287/

The reason being that, since we're a strictly informational site, there's really no reason for us to play the SEO game. Considering the nature of names and album titles, the generated title slugs would actually end up making the URLs dirtier. Having shorter, ID-based URLs also makes it a lot simpler to manually construct links, too.

[Secret Squirrel]

If that's the case, why change them at all? We've already got thousands of links pointing at our content.

And I agree. We've got great search-engine presence without playing any tricks, so it's probably not worth it.

[Liontamer]

No pressure implied, but whatever happens to finalize things with all of the URLs, I'll be glad so we can finally start linking things at OC ReMix. :-)

[R-T]

If you want to drop the album title part, how about reducing URLs to http://vgmdb.net/a272 for albums
http://vgmdb.net/m112 for artists (musicians?)

[mrnutz]

I can understand the desire to move to a cleaner format. However, this implies supporting backwards compatibility (the redirects) forever doesn't it? That seems like extra work. Are there other advantages?

[Gigablah]

The redirects can be done with 2-3 lines of code in .htaccess, as long as we don't include stuff like the album title in the URL (otherwise we'd have to do it via the script). The other advantages (apart from short, clean addresses) would probably only make sense to a web developer (abstraction, masking, etc).

R-T's proposed format is certainly attractive (and it's currently in use at vndb.org).

[R-T]

R-T's proposed format is certainly attractive (and it's currently in use at vndb.org).

..who borrowed the idea from Anidb.net.