I saw another image and had a new idea for Img.Gvid.TV:
Maybe this is easy, maybe not.
As on WikiMedia Commons, they list the pages that image is used on.
If Img.Gvid.TV had that, we could see an image and follow it to the forum if posted/commented to see or add takes on it.
It might draw to the forum folks who just post images.
Sure it might be nice to have voting, comments, and metatags on images to better organize them - but at this stage in Goat/Gvid's growth it's overkill and back-linking seems like a good feature for now.
I've thought about that. It was one of the first gvid features, so I've done it before. It's death by 1000 tiny cuts for why it ends up sucking.
Referrer headers not being what they used to be (which makes sense for privacy). Most request coming with none. People browsing on mobile apps that inject an id for the mobile app instead. The ambiguity between people who came from the root page of one domain, vs their browser just cutting off anything but the host. Low ability to predict if the resulting URL will be useful for people to click on or land people on an irrelevent page. The web no longer being a web means urls don't map to specific useful pages as consistently as they used to.
The result is you basically have to log the stuff, and then comb through it to detect url forms worth white listing. I've done exactly that on gvid. But that white listing phase just never happens. The white listing phase is necissary because the referrer data you get (what page a user was on just before landing on yours), is 99% junk.
A lot of site owners don't want to link out in general. I do. But it ends up being a pretty bad experience just to do something nice for other web admins. I'll do something nice for the sake of other web admins. I'm just not going to give myself 1000 papercuts to do it.
Almost better would be to have a text box that asks a user "where did you come from" and then coroborate that with referrer data. You'd actually get a better result just blindly accepting whatever a human says because 10% accuracy beats 1% accuracy.
I'm going to re-add the idea to my todo. It's been there before.
If I understand what you're saying above, you meant that you'd track-back to usage of the image to other sites all over the web.
I think I may not have been clear. I only meant track-back to use within the Gvid and SubMatrix network (easier as you control the sites, their code, DBs, search, etc.). Perhaps add other friendly domains (ie. Projex, WikiSpooks, etc.) if practical and useful.
This is an example what I was talking about, at the bottom under "File usage" - this image is used on four pages.
I think this is kind of related to what I think you were talking about: https://projex.wiki/wiki/Fringe_Deplorables_Web_Ring
If you coded it, IMO it wouldn't be unwarranted, in fact it's earned, in the corner of the "ad", to take credit and feature "SubMatrix" linking back to a page/article about how you coded it, what it is, who's on the list, how to earn a spot on the list, etc. - and importantly you'd have to clearly and concisely outline the expectations of SubMatrix users (we don't want riffraff), assuming the article explaining the "ad" invites them to make an account.
Oh yeah. I have also put in that idea recently. Gvid gets no comments on it's site at all. The image host doesn't even have comments.
So the broader idea is to turn matrix into the comments provider for both. And that would effectively also link back to where a discussion(s) is happening.
And my idea was if matrix doesn't yet have a comment thread for the link, instead of displaying a comment box, have a cheeky "enable comments by posting to submatrix" link. Someone had the bright idea of suggesting I support bookmarklets targetting the submission form, so it would leverage that some.
If we wanted to incorporate projex, maybe we would use an API call from your site to mine, to self-report. That or have a UI you can add it manually. I wouldn't trust the referrer data even if it pointed to your site because it goes though people's browsers and it's shocking how random the data gets. It would end up pointing to wrong pages regularly.
I love the idea of bookmarklets, tags, and notes. I love to keep galleries of my fave memes in topical categories. Wish I had infinite time to only fuck around.
Log in
*fake and gay (secretly or not)
All of "peaceful" Trumps promises were lies and it's clear Trump regularly bends over for Bibi.
I saw another image and had a new idea for Img.Gvid.TV:
Maybe this is easy, maybe not.
As on WikiMedia Commons, they list the pages that image is used on.
If Img.Gvid.TV had that, we could see an image and follow it to the forum if posted/commented to see or add takes on it.
It might draw to the forum folks who just post images.
Sure it might be nice to have voting, comments, and metatags on images to better organize them - but at this stage in Goat/Gvid's growth it's overkill and back-linking seems like a good feature for now.
@x0x07
I've thought about that. It was one of the first gvid features, so I've done it before. It's death by 1000 tiny cuts for why it ends up sucking.
Referrer headers not being what they used to be (which makes sense for privacy). Most request coming with none. People browsing on mobile apps that inject an id for the mobile app instead. The ambiguity between people who came from the root page of one domain, vs their browser just cutting off anything but the host. Low ability to predict if the resulting URL will be useful for people to click on or land people on an irrelevent page. The web no longer being a web means urls don't map to specific useful pages as consistently as they used to.
The result is you basically have to log the stuff, and then comb through it to detect url forms worth white listing. I've done exactly that on gvid. But that white listing phase just never happens. The white listing phase is necissary because the referrer data you get (what page a user was on just before landing on yours), is 99% junk.
A lot of site owners don't want to link out in general. I do. But it ends up being a pretty bad experience just to do something nice for other web admins. I'll do something nice for the sake of other web admins. I'm just not going to give myself 1000 papercuts to do it.
Almost better would be to have a text box that asks a user "where did you come from" and then coroborate that with referrer data. You'd actually get a better result just blindly accepting whatever a human says because 10% accuracy beats 1% accuracy.
I'm going to re-add the idea to my todo. It's been there before.
If I understand what you're saying above, you meant that you'd track-back to usage of the image to other sites all over the web.
I think I may not have been clear. I only meant track-back to use within the Gvid and SubMatrix network (easier as you control the sites, their code, DBs, search, etc.). Perhaps add other friendly domains (ie. Projex, WikiSpooks, etc.) if practical and useful.
This is an example what I was talking about, at the bottom under "File usage" - this image is used on four pages.
I think this is kind of related to what I think you were talking about: https://projex.wiki/wiki/Fringe_Deplorables_Web_Ring
If you coded it, IMO it wouldn't be unwarranted, in fact it's earned, in the corner of the "ad", to take credit and feature "SubMatrix" linking back to a page/article about how you coded it, what it is, who's on the list, how to earn a spot on the list, etc. - and importantly you'd have to clearly and concisely outline the expectations of SubMatrix users (we don't want riffraff), assuming the article explaining the "ad" invites them to make an account.
Oh yeah. I have also put in that idea recently. Gvid gets no comments on it's site at all. The image host doesn't even have comments.
So the broader idea is to turn matrix into the comments provider for both. And that would effectively also link back to where a discussion(s) is happening.
And my idea was if matrix doesn't yet have a comment thread for the link, instead of displaying a comment box, have a cheeky "enable comments by posting to submatrix" link. Someone had the bright idea of suggesting I support bookmarklets targetting the submission form, so it would leverage that some.
If we wanted to incorporate projex, maybe we would use an API call from your site to mine, to self-report. That or have a UI you can add it manually. I wouldn't trust the referrer data even if it pointed to your site because it goes though people's browsers and it's shocking how random the data gets. It would end up pointing to wrong pages regularly.
All good ideas.
I love the idea of bookmarklets, tags, and notes. I love to keep galleries of my fave memes in topical categories. Wish I had infinite time to only fuck around.