Gossamer Forum
Home : Products : Gossamer Links : Discussions :

Filtering Poor Quality Submissions

Quote Reply
Filtering Poor Quality Submissions
Hi,

After upgrading from 1.13 to 2.1.1 one of the features I've been most impressed with is the Database > Links > Delete search facility.

In 1.13 you couldn't search through waiting submissions and delete the ones where people have obviously ignored your submission criteria or preferred formatting.

I've found that I can 'filter out' (delete) site submission en-mass that don't fit with my preferred format, e.g.;

• Title contains additional or extraneous wording: " - ", " : ", "...", etc.
• URL contains affilite links, etc.: "/?", "id=", "vstore"

Whilst some of these may be from novices and it is a known bug-bear of operating a directory that not everyone will head your advice, I was wondering if there was a way this could be hacked into the actual submission form to make it clear at the point of submission?

I was thinking that you could add a series of 'unwanted' snippets to be checked against each field when someone submits a site. Then, for each 'fault' you could display a reason why it hasn't been accepted and offer the opportunity to correct it.

My hope is that eventually enough people will be 'taught' how I prefer submissions and it would lead to a marignal decline in ones that didn't fit my preferred format.

I'm asking because I think it would be an additional feature that a lot of people could take advantage of. For example, filtering out 'freepage' sites, affiliate links, untidy submissions, and maybe filtering in specific country based domains e.g. "co.jp", ".co.nl", etc.

Am I just being lazy, or is this something that you think would be of benefit to everyone?

Also, do you have any 'filters' of your own that you use to catch unwanted submissions?

All the best
Shaun
Quote Reply
Re: [qango] Filtering Poor Quality Submissions In reply to
Quote:
URL contains affilite links, etc.: "/?", "id=", "vstore"

For my own directory, such a policy would exclude a number of sites with a validly dynamic structure... Just something for everyone to keep in mind, as each directory owner knows the nature of his or her content the best. :)

Dan
Quote Reply
Re: [Dan Kaplan] Filtering Poor Quality Submissions In reply to
Dan,

Thanks for your input. I appreciate that everyone would have different requirements, that's why I was wondering if it would be possible to create some configurable filters - then each directory owner could tailor it to their own requirements or preferred format.

Maybe I should look at doing this for my own directory first? If I can make it work I could post the mod for those who might want to use it Smile

I'll think it through a bit and then see what I can come up with.

All the best
Shaun
Quote Reply
Re: [qango] Filtering Poor Quality Submissions In reply to
I quite like your idea Shaun. It could almost be a seperate setup page from which you could select pre-set "criteria" which all submitted links must fit within. In there, for the more perl savy, it could also have a field to input a regex which would ultimately override anything else. I know in the database fields you can currently enter regex's, but just being able to click a check box to choose some criteria would be just the business.

It could be as basic as max and min title length, description length, some url formatting for situations like yours, and country suffix's would be a good feature too. I use AdvertPRO which has something similar when you're selecting how and where banners will be displayed - ie. which browsers, which countries etc. It definately makes setting it up a breeze - and very quick.

r.
Quote Reply
Re: [ryel01] Filtering Poor Quality Submissions In reply to
I'm on the opinion, that it's fine to filter somehow the submissions, like stopping automatic submissions with referer checking & duplicate submissions with dupe checking. But better to let the users decide what they like.

Further solutions:
- Manual solution: checking link by admin manually, and decide if they are good for the site or not (that's how currently works)
- Automatic solution: allow the visitors to decide what links they like better by creating reviews & voting to links. Links are awaiting in a public quarantine, where users can vote for them. Only those will get into real database, which was checked, reviewed, voted by users, and reached the minimum requirements. Then delete old links, which did not meet minimum requirements, after 30-60-90 days...

What do you think about automatic solution?

Best regards,
Webmaster33


Paid Support
from Webmaster33. Expert in Perl programming & Gossamer Threads applications. (click here for prices)
Webmaster33's products (upd.2004.09.26) | Private message | Contact me | Was my post helpful? Donate my help...

Last edited by:

webmaster33: Sep 25, 2002, 2:49 AM
Quote Reply
Re: [webmaster33] Filtering Poor Quality Submissions In reply to
I think you've possibly misread the original post - I'm not talking about filtering 'sites' themselves, rather the data that is entered in the "submit a site" form.

A common example I come across is dashes in the site title, e.g.; "Jims Tackle Shop - Get all your fishing gear here!"

My personal preference is to have just the title - I don't really like it when people pad-out their title with extra wording, I prefer them to put that stuff in the description. Ideally the above example should just read; "Jims Tackle Shop"

What I'd like to do is have the submit form CHECK for this sort of thing then display a polite notice saying something like: "Your TITLE includes a dash (-) which does not fit with our preferred format, please check our site submission information and try again."

That's just one example. Since everyone operates their site differently a series of optional custom filters could be tailored by each owner to suit his/her requirements or personal preferences.

The basic idea is to politely inform the visitor about how you prefer them to format their submission data and to weed out poor formatting before they submit their site.

Hey, maybe I'm just being a bit too fenickety about formatting? Smile

All the best
Shaun
Quote Reply
Re: [ryel01] Filtering Poor Quality Submissions In reply to
ryel01,

I like the idea of being able to set min. max. lengths for fields. Maybe that could be an 'add' global that is checked in add.cgi?

I know you can set it on the form for an input field, but textarea fields don't have a maxlength so it would help avoid overly long descriptions.

I think I'll have a play around with these ideas over the coming weeks and see what I can turn up Smile

All the best
Shaun
Quote Reply
Re: [qango] Filtering Poor Quality Submissions In reply to
You can give guidelines to the users, so they should not put extra info in the Title, but you can not force them, until you do not modify manually.
If a user see there is not allowed to post "Jims Tackle Shop - Get all your fishing gear here!", then he will post "Jims Tackle Shop. Get all your fishing gear here!", or using any other delimiter character. Even space, which you can not to filter out...
EDIT: CORRECTLY: space, which you may not want to filter out... (was misunderstood, so this is how I meant)

Note, 10-30% of your users (if not more), will not do, what you would want. They always try & find new tricks to avoid dirturbing rules... Wink

Best regards,
Webmaster33


Paid Support
from Webmaster33. Expert in Perl programming & Gossamer Threads applications. (click here for prices)
Webmaster33's products (upd.2004.09.26) | Private message | Contact me | Was my post helpful? Donate my help...

Last edited by:

webmaster33: Sep 25, 2002, 4:27 AM
Quote Reply
Re: [webmaster33] Filtering Poor Quality Submissions In reply to
Quote:
If a user see there is not allowed to post "Jims Tackle Shop - Get all your fishing gear here!", then he will post "Jims Tackle Shop. Get all your fishing gear here!", or using any other delimiter character, even space, which you can not filter out...

Isn't that the point?

You can filter spaces in any case.
Quote Reply
Re: [webmaster33] Filtering Poor Quality Submissions In reply to
I see what you're saying, and yes, this will only apply in a few cases. Most people DO take the time to read my guidelines and DO make a good quality submission.

You're also correct in that some people will always find a way around any 'checks' or 'rules' I impose and I know there is nothing I can do to stop them.

I suppose I'm looking to put something in place to give friendly advice about how I'd prefer submissions to be made.

Perhaps this is something I should do for myself - I just thought I'd share my ideas to see if they were of interest to anyone else Smile

All the best
Shaun
Quote Reply
Re: [Paul] Filtering Poor Quality Submissions In reply to
Khm. Well my mistake. Was misunderstandable.
Correctly I meant: space, which you may not want to filter out...

Best regards,
Webmaster33


Paid Support
from Webmaster33. Expert in Perl programming & Gossamer Threads applications. (click here for prices)
Webmaster33's products (upd.2004.09.26) | Private message | Contact me | Was my post helpful? Donate my help...
Quote Reply
Re: [qango] Filtering Poor Quality Submissions In reply to
If you hook site_html_add_link you can probably do something like this:

Code:
my $input = $IN->get_hash;
my @fields = qw/Title/;
my $chars = GT::Plugins->get_plugin_user_cfg('Your_Plugin')->{Filter};
my @bad = ();

foreach my $field (keys %$input) {
next unless grep $field, @fields;
if ($input->{$field} =~ /\Q$chars\E/) {
push @bad, "$field contains invalid characters!";
}
}

That's just a quick code sample that will probably need altering but it is just a quick idea for you. The plugin would have a user option called "Filter" which you'd then use to enter your blocked characters.
Quote Reply
Re: [Paul] Filtering Poor Quality Submissions In reply to
Sorry user_add_link is the hook.
Quote Reply
Re: [Paul] Filtering Poor Quality Submissions In reply to
Paul,

Thanks for the code, I'll have a go at whipping something up when I get some time Smile

All the best
Shaun