CI "must haves"

[eluser]Narkboy[/eluser]
Most of my work is on secured intranets. Search terms are most frequently drawn from preset lists which change very regulaly. A bookmarked search is more likely to be invalid than not, which frustrates users more than having to redo the search - think drill-down reports and such. It also means that they are outside the normal usage patterns, which can trigger intrusion detection and get me out of bed for no good reason.

For public or low-security searches, or searches with a longer life, I tend to use segments. Mostly because now I'm used to them, I prefer them to query strings. You can get pretty creative with them without making the url look bad:

http://example.com/search/age/6/color/bl...r/whatever

Can be dealt with just as easily as:

http://example.com/search.php?age=6&colo...r=whatever

But it looks a lot nicer, and is easier to read. You can also search from the address bar rather than the form if you like. All you have to o is ensure that segment[n] is a var and segment[n+1] is a value.

But yes - rereading my post, it does seem odd that I don't want searches bookmarked. I'm so used to working on private, heavily secured and logged sites that I forget how real sites work! I'm allowed to be soooo much meaner to my users that on public sites!

[eluser]pbreit[/eluser]
Yeah you definitely don't work on real sites. Thats the dumbest URL vie seen in awhile. Thank goodness not what we see in the real world. Drill down reports are an obvious candidate for query strings. I guess all your projects are solo efforts as well?

[eluser]Narkboy[/eluser]
Wow - claws away Miss Kitty. Exactly how is a URL dumb? I mean - aside from being overly long and requiring users to actually understand URLs, which is exactly what query strings do.

Duimb, in any design setting, should be defined as more complex than is requried. Good design is about taking things away, not adding them. What does a question mark and an ampersand add to a url, exactly?

You are clearly far more experienced than me, more wise, more powerful. Forgive me - I should not have spoken on Your Divine Presence. I shall not repeat the mistake, please forgive me.

But since I have intruded upon your precious and worthy time, perhap you can enlighten me as to the precise difference, in terms of 'dumbness', between the first URL and the second?

[eluser]pbreit[/eluser]
You are forgiven. It's just irritating when people come in and defend such a lame aspect of the framework. Queryatrings are a core part of HTTP/HTML and are the correct technique to use for parameters. Having to workaround CI's limitation on such a basic aspect of web development is annoying.

Ps: the second URL has the obvious benefit of the keys and values being explicit. Further, it can be called from a form. I challenge you to find another search or reporting drill down function that doesn't use query strings.

[eluser]Narkboy[/eluser]
[quote author="pbreit" date="1290467347"]You are forgiven. It's just irritating when people come in and defend such a lame aspect of the framework. Queryatrings are a core part of HTTP/HTML and are the correct technique to use for parameters. Having to workaround CI's limitation on such a basic aspect of web development is annoying.

Ps: the second URL has the obvious benefit of the keys and values being explicit. Further, it can be called from a form. I challenge you to find another search or reporting drill down function that doesn't use query strings.[/quote]

Thank you - your forgivness means a great deal to me.

Honestly, if you have an issue with something as fundamental as the url approch taken by a framework, surely you'd be better off using something else? There are a number of frameworks that work better in cases where you need query strings extensively.

Personally, I've never liked the ? = & approach; it's ugly. It does have benefits, and yes, explicit keys is one, user familiarity (among some user groups) is also avery big plus. I'd rather not use it though, which is why I've spent so much time using CI - it removes the issue for me.

I do have problems reverting back, and an upcoming PayPal project has me unsettled because it seems I'll have no choice. Not looked too deeply, but it seems it's query strings or bust. So far, I've managed to do all external tie-ins via api / soap or similar. For me, it's better to handle the mechanics out of site of the user, though there are excetptions.

Ultimatly, for every benefit, you pay a cost. So far, the cost of simple urls has been very low for me, so I'm happy. My clients agree, as do the other people I work with. I'll figure the rest out when I reun into it!

As to finding another system that dosen;'t use query strings? Why would I? Mine dosen't, and works all the better for it - at least, that's what the sales ledger tells me. I'm more than happy to conceed that I may find an issue with it the requires qs, and I won't deny there have been days.... But - that's coding