Quality Standards in The Library

I know we're not all on the same level of filter building, I know we all have different backgrounds in our ideas, and I understand we're all different. Yes yes, blah blah blah.... I hate talking about this because it always seems unfair, but I have to do it. I've been working professionally for several years now, and in doing so I come across practices in art which should be dealt with. So this is basically a list of *bit harsh* critique on filters in general. I've been using these as standards myself. I don't want to put people down, my intentions are more concerned with the overall filter, decision, and the final output, and nothing personal.

Warning: super long post!

Filters submitted to the public library should not...

1. have controls that go out of whack and break visual style when a value is changed.

What: you download a filter that produces some nice lava texture. The presets look nice and so you expect it to look nice. The first preset is ok, but the moment you slide 1 value on some slider, the whole lava texture vanishes, the output is green and it leaves you with a mess. (Extreme example!)

Reason: controls have no limit or remapping. By remapping a control, you're limiting the filter's range of capabilities, but you also give the opportunity for the user to explore your filter without going "what just happened?!"

2. have tons and tons of versions that do almost the same thing!

What: You go to the library, see 7 versions of the lava texture all submitted as a separate filter. The author claims that they're different but they're visually the same and share little difference between each other.

Reason this isn't so great: If you design a filter well, it should be able to stand on its two legs. You really don't need to add more legs. Aside the humor, this is a flaw in design. A filter doesn't need so many siblings (versions.) The variations should be brought upon by adding controls and variables and never through 5 versions of the filter. I hate to see myself hopping around from 1 version of the filter to another just so I can get some aspect. I hate wasting my time looking through 5 different filters for some attribute that may get me what I want. Other than that, for every new version added it really just takes away from the original idea, its distracting to look through each version, and it really doesn't add to the library's wealth and goodness. Not everybody has a copy of filterforge professional. Do remind yourself that there are people who cannot edit filters so having more than 5 versions is trouble for those folks who just want results.

There are some exceptions to this of course.

Solution: Make one, build on it. Add controls and design variables to give the filter more usability and substance. Don't add more to the filter library.

3. have a boat load of features and it can make everything!

What: Author decides to build photoshop in filterforge. Or... say I download a lava filter, then I change the slider and it becomes water, then it becomes a pig, then it becomes an abstract shape! Oh each value has more controls! Like 100 controls! Oh boy it must be a useful filter! No. These filters sink really quickly and it sucks using them.

Solution: don't make them.

4. serve a really really really really really specific purpose.

What: Filter produces lava. Very cool. Filter produces lava with hearts and candy all over it. Not cool. Author made it for Christmas to surprise the kids.

Solution: Don't upload super specific filters. Make them general so everybody else can enjoy the filter.

5. be a cliche.

What: Here's another generic wood texture! Oh boy I wanted that so badly! end of sarcasm.

Reason why this is bad: After 100s of lovely wood generators it sounds like we'd probably want less of them. While each filter has some purpose, we probably don't need 100 more filters that serve similar trades. On the other hand if it looks visually different, produces a unique result then it becomes worth uploading.

6. be crappy.

What: Not all of us are designers, but we know what bad things look like, what "good enough" is to "excellence." We have a list of featured filters, we have popular filters. So there are "marks of quality" that people look for in a filter. A crappy filter ignores all of those marks and decides to be just its own. Of course, a first time filter is always going to fall into some level of "not super awesome," but that doesn't make the filter "crappy."

A crappy filter is defined as something that basically just doesn't do anything great. It produces bad results no matter what, does nothing, it is boring, doesn't offer any substance, it's a waste of time, wastes your time using it, it's troubling to use, it lacks any technical skill and in fact ignores all of that, doesn't have any visual element, follows no standard, and the author doesn't try to improve it so it just rests in the library being stupid.

Solution: none!

I wrote this over a period of a couple days. It's a growing list. So you know... it's long Yeah sorry. Add your own!

Excellent post Skybase! Let me add a few of my own observations.

You discovered Surface Mode..
Yes, it is indeed very cool, but so did the 5000 other HU hunters think before you. Surface Mode bling bling only impress a few new visitors, so reconsider before you submit your "Totally Unique Goo Liquid Chocolate Metal Generator" filter. We have alot already

Your filter is slow..
Rather than not, assume that your filter is too slow. Before you submit, make sure you test your filter with different images. If the filter speed bothers you slightly, chances are others will find it unacceptable. Bye bye HU points..

Ask yourself: "Is that 22th. blur/sharpen/highpass/minimum/median/maximum component really necessary?", "Can I skip some steps by redesigning for performace?". Often a filter is build gradually with visual result in mind, not performance. Nothing wrong with that, but remember to give optimization a go before submitting (and ask in the forums if you're stuck).

Filters submitted to the public library should not...

continuing from the commentary...

9. Be ridiculously simple.

What: A filter that generates cards of color. You can choose between 5 colors and it will put cards everywhere. A dot generator! Makes circles in 2 colors! Noise generator! This one makes noise that looks mildly like perlin noise! Gradients! Everywhere!

Whhaaat? Simple filters exist, and there are tons in the online library. It's not a bad thing, it's just there are ones that seem ridiculously simple. 2 Colors on top of each other with a mask on it? That's it? Oh come on, thanks for wasting my precious filter browsing time! (lolol) but you see, the point here is that the filter may produce good stuff (and sometimes simple does produce good stuff) it shouldn't be to a degree wh ere I'm just wondering how it managed to get on the library

Solution: Add more features and put more thought into it. Really.

10. be the ugly duckling.

What: the ugly duckling became a beautiful swan in the end. Sucks but it doesn't work like that in reality. What's ugly is ugly in art, if you want it that way, it'll remain that way. A filter that produces a grayish pink green bubble thing doesn't go too far until the author deals with the colors and perhaps the properties. Learn the principles of design, color, and other standards that make things look great. They're simple, but they take practice to understand. Learn what makes a good filter good and practice them. You don't have to know all the negativities, you just have to know what makes good stuff good.

I mean there are simpler solutions to maintain content in the library, but most of it is up to the discretion of the filter author. We do have high usage stats and stuff which does work as a rating system, it's just not actually reflecting on what people prefer to see. Voting I guess is the easiest, but there's something about it that doesn't work with the Filter library.

And see, awards become a matter of opinion and interest between a couple people. We can easily skew opinion especially in a small group of niche users. The editor's pick filters is already "good stuff" which is what I'd rather see. But there's nothing wrong with producing a blog article or two about which filters do wonderful things. (Honestly think its cheesy to have batches and stuff. haha.)

So it ends up becoming up to the user's discretion to submit good stuff really. I'd produce tutorials which cover basic and advanced topics and that'd be a way to bring a level to people who just began. We do have a wiki page also, it just doesn't find itself being exposed as much.

you dont need quality standards in the library. there is a good reason why it all gets lumped in together... several reasons, actually. a new filter maker is going to very often make junk. he's not going to be familiar with what's in the library all that much and he's going to make mistakes... just like we all did. i say let him. one needs to output something to get started. if you do nothing but reject and invalidate new works, even if they are crap, you'll never develop as many good filter makers. and that's just the way that is.

however, what could be done, is a new, high quality library. it could even be done with a minimal output fr om FF, inc. if filters had a rating and at a certain rating they automatically got copied over into the HQ library, you'd have what you want. you could even make that rating the HU (high use) rating. a filter gets to HU and it gets copied over into the better library. mind you, i said 'copied over', for you still leave a copy of the HU filter in the main libary.

you might, particularly some of the older members on the forums here, have noticed that we're not seeing very many FF, inc. employees on the forums these days. GMM is around, but where are the others? bella seems to have been gone for quite a while. i dont see onyx around and even vlad has been missing for a while. and this is why i said there is more than one reason for not messing with the libary. someone has to do the work and i dont anyone in here volunteering. anyone could start a high quality library web page. wouldnt be that hard, but i dont see any volunteers and the reason i point out the lack of FF, inc. employees is that i think the ranks have dwindled. my guess is wh ere there used to be 5 or 6, there may now only be 3 or 4, which means that work done at FF, inc. is being done by fewer people, meaning jobs, projects and priorities have to be administered more tightly and there probably isnt a lot of extraneous time for lower priority items, like revamping the entire library.

and as for some of the limits some of you want to put on the library, ppppphhhhtttt!!! no way, jose'! i love filters with controls that 'go out of whack' and some of the long rendering ones. tons of versions, yeah, i dont need six parquet wood flooring filters, but at the same time, if you squash this aspect of filter submitting, you likely squash that one guy that comes along and makes that one wood parquet filter that blows the other six out of the water. i specialize in making filters that make lava and also water and then something else. you dont see many of them in the library because i hoard these... they are my favorite filters. they are inspirational! i click on 'next variation' and i get a whole new look i've never seen before! this is one of the most marvelous features of FF, it's infinite variety and it's ability to inspire! if you ever want to squash creativity, put start putting limits on it and you'll get there. add in a few arbitaries while you're at it and poof, creativity hits the rut.

i'll tell you a little secret about creativity. it's meant to communicate. you're trying to get noticed, to tell a story or inspire or create an emotional reaction... you're trying to communicate something. when you get no reaction, no response at all to the thing you're trying to communicate, that in itself is a communication and it shld bloody well tell you you need to either promote more or go in a different direction. i've done lots of work in various forums across the web and whenever i fail to produce a response, i either abandon the work or revise it or promote it better, but normally it tells me it wasnt good enough. if i take that no response and go all disheartened and disenchanted and quit, i've made the wrong decision. a young filter author here shld take things the same way; produce whatever seems good enough to appeal to someone. if it gets no response within a period of time, revise it and re-submit or comment on your own filter and ask questions. the wrong thing to do is to disqualify or reject the newbie filters. they'll quit and not come back and that would be a loss to the whole community.

i download every single filter in the library. when i start up FF it takes quite a bit longer to load because of this, but i find it well worth it. one single crappy filter might well contain a routine or an aspect i've never thought of before and it might well inspire me in making something of my own. i dont like limits. i like infinite possibilities. you guys shld have been here in the early days when filters could only have 10 controls max. that one drove me crazy. one of my filters has over 120 controls and it's one of my favorites. it constantly inspires me.

so, be careful what you ask for, you may get it

Very Well Written Craig.. and I do agree with you 100%...
I remember that discussion of the two different types of library..I thought that this was a good ideal. However not so much the paying for the better filters part.....
I used to belong to a site that was community driven creating UIs for a gaming platform...Real early 2000s...They had a mix of submitted material the way that the FF library is set up.....Eventually they had to organize the submitted material because they were losing people not wanting to dive through all the not so good stuff....

Remember the beginning when there was actual intervention from FF when filters were submitted..
hehe...Uber....There will be dilla as the Official "Filter Forge Seer"....Then Uberzev as the Official Filter Forge Over-Seer...

This is a good thread.
If the community could produce a unified list of guidelines that all expert filter makers will consider appropriate, I will add a link to this list at the guidelines page.

Craig, I partially agree with what you're saying. I just don't think its as applicable when you have thousands of people looking through the library, downloading the filters and using them.

Here on FilterForge forums we tend to forget the masses of folks who browse and download through the filterforge library and use the filters we make. We don't know who they are, all we know are some numbers and usage stats. So if anything, if I upload a filter with some control that goes whack mode, the users who download and use my filter may not understand that.

Remember, not everybody has a copy of FilterForge profesional. There's no way for some people to fix filters if there are issues.

Users expect filters to behave the way they see in the previews and description. If the filter doesn't behave that way, they expect it, then it's just a filter that doesn't work. At the base level people just want to see stuff work. On the other level, people want to see amazing things. But I'm talking more on the base level because we can build fr om the base and get amazing stuff.

So you know, it's nice for us to have this comfortable awesome environment to chit chat do whatever, but I just don't want to forget those people who actually use the program for all sorts of purposes. We're pretty much a happy family here. I'm surprised a software forum can even get so busy. Just check out all the other graphics related software forum. The forum's just a barren landscape of a couple questions. So I wouldn't want the super strict lim its to be law, I just want them to be something to follow.

On another note, my Dabber filter hit 6000+ downloads recently. I don't know those thousands of people who use my stuff, but as a way of saying thanks to all those who actually cared to try my stuff out, I want to continue making good stuff.

skybase,

thanks for responding to my response. i was mainly directing it at your first post here. i was a bit strong with it because i've seen folks try and 'police' creativity before and it ALWAYS ends in less creativity! now, i doubt you were trying to be the filter police but, i've just seen this too many times before. you NEED that one, common, anything goes, library! it's a MUST! the newbie MUST be free to post up the worst possible crap in the world without restrictions, except for blatant and obious plagarism.

now, i dont mind if you put a rating system on the filters such that one could do a search based on those ratings. that's fine. i dont mind tiers of libraries, pay-for or not. that's also fine. i dont mind helping newbies, or oldsters, for that matter with some sort of system, AS LONG AS YOU HAVE THAT ONE, UNRESTRICTED LIBRARY THAT ANYONE CAN POST TO WITHOUT PREJUDICE!

i dont think the current library is the problem here, anyways. i think it's the search and sort of said library that's the problem. it's also the problem in the program itself, but i'll not re-hash that part of things right now. a public library with books has thousands and thousands of books within it, yet one can relatively easily find what one is looking for because the categorization of said thousands is done up nicely and standardly regardless of what library you are in. so, i would put it to everyone here that the problem isnt how many filters there are in the library, but rather sorting through the library and finding that gem you are looking for. you could have 100,000 filters in there and 90% were crap and have it not be a problem if you had an easy and reliable way of finding those few filters you really want. so, i see the current discussion as being a bit off-target. i say it's not the volume of filters or the quality of filters that is the problem; it's the ease and efficiency of finding what you want that is key to all this, no?

Craig, as far as philosophy goes, I think the current almost-free-for-all library environment is great. It works, there's a load of fun stuff here that usually isn't practical but we may find it adorable. And yes I'd want to see some organization in the library so we can get to stuff we want faster. But this isn't too much about the library itself, it's about what gets submitted to the library.

But let me be clear, what I don't want is super policing the library. What I would rather see is people policing themselves, catching issues in their own filters when they see it. Of course, on the higher end, people should eventually practice design principles to produce better designs.

So after all these years of sticking around in the forums, I kinda get to know you folks. So I do understand workflows and what goes into making a good filter, but most others don't know that. They don't care if you worked hard on a filter or not, users would rather see your filter work in their favor than anything. I'm sure talking about your creative process would give insight into what goes into making stuff like this, and I'm sure people will appreciate that. But most would rather see things working. That's the consequence of exposing your work publicly, people will judge work no matter how hard you try to explain.

Unlike books filters are like little programs that have output. We easily get distracted by the overall resulting artwork produced by the filter, but what we're really making is functionality for that art to come out right. The standards here isn't so much to do with the art itself, it's about making the filter work better and getting the work to come out right. It doesn't happen over night, nor does it happen because you upload tons of filters onto the library.

And speaking on the part of creativity, we're already restricted by FilterForge's functionality. While we get more function for every new version, you do have to remind yourself what was built with version 1.0. All those were darn crazy for what the program was capable of back then. Free for all or not, I'm sure you can get creative anyway.

... now I wonder if we should start compiling a condensed list instead of some tl; dr posts.

way too much tl;dr here

Someone write a proposal with bullets and formatting that doesn't make my head hurt.

You might not like the elitism Gene but the filter library needs more "Breaking Bad" and less "Jersey Shore". IE filter diarrhea.

Quote
FF is just PhotoShop Actions with a GUI

And Windows is just DOS with a GUI.

Perhaps we could encourage people to post to the user gallery for feedback before deciding their filter was worth submitting.

The submission wizard could even have an automated process for doing this.

hi skybase,

i do understand what you're saying. helping authors make better filters is always good. i probably didnt read your original post closely enough. please understand i never said any of this in anger or am i upset with anyone here at all. i know emotion doesnt always translate well in text only. your ideas are fine and good. i just wanted to make sure we werent rabble rousing here for 'filter police'. like i said, i've seen it before. so, i might have lumped you into that category without truly duplicating what it was you were after. and even if i did, i probably still would have commented with a warning. i think i may have backed you into a corner needlessly, and for that, it wasnt intended. i'm always pro-artist, you included and as for people policing themselves, i'm all in favor of, providing they dont stifle themselves into non-action as an artist (it does happen, trust me).

i do know there are helps here in the forums, in the wiki stuff, within the program and so on. perhaps there shld be a forum specifically dedicated to helping newbies out. yes, i know we have the 'creating filters' forum, and that's good, but after so many pages and pages of stuff, it gets harder and harder to find things, such is the nature of forums. if i'm a newbie coming here looking for help and 'tips', i'd be pretty overwhelmed. the program itself is a bit overwhelming at first, particularly when you start playing with the editor. vladimir and company have done a remarkable job in keeping it as simple as possible, but the program has gotten quite big and there's a lot to it these days. thus, helping the newbie or even the oldster, can be a bit of a trick.

so, how do you help that new guy build a better mousetrap? some folks make filters and never join the forums. ok, hard to help them. you set up pointers to helps and hope they get there. but i've seen a number of newbies come into our forums and they get razzed or scoffed at for making 'crappy' filters. they leave and dont come back. so, i'd say your first point of help has to be creating a safe environment for the creative... even if the creative is 'crap' and that was the main point of my responses in this thread. that a four year old scribbles something with a crayon on a piece of paper is remarkable. that it might look like anything is even more remarkable. so momma comes along and goes 'what the hell is that mess?' and the four year old is crushed. he was making, in his mind, a beautiful rendering of momma herself and was presenting it to her as a gift of his own making. had momma taken that same 'picasso' and praised the child for the effort and gift and even hung it up on the fridge as the masterwork it truly was... the child beams and tries another.

so, i probably over-reacted to your posts thinking 'oh oh, here comes a stomp on creativity in the form of rules and such'. in looking back over the posts i could have tempered that a bit.

so, how do you get folks to read and follow your 'design practices'? and how do you make a set that is open enough to the creative but tight enough to produce good works, because i still disagree with a couple of the guidelines. like i said above, some of my 'best' filters (read: favorite) are wide open. they do produce 'everything'. they are meant to. some of quite specific, especially the texture type and i use remapping to keep them 'tight'. but i'm currently working on another which is a 'forge' or 'factory' type, which is meant to have a VERY broad range with very little remapping. i'm the guy that pretty much talked vladimir into removing the 10 control limit. i like 'broad' filters... well, some. but you are right in asking for some control there. i've certainly seen the ones where you have 9 presets that were pre-rendered and those are the only workable renditions you're ever going to get out of that filter. so, what's your solution? how do you get folks to apply 'design practices'?

As I thought about this, for the User's Choice, perhaps I could propose something which I did in a similar situation over a decade ago.

Rather than a simple "Thumbs Up" kind of thing, my first version of a voting system had users given a limited number of "tokens" they could use to flag entries. They received a specific number of gold, silver, and bronze stars.

If you had given out all your stars and a new entry appeared and you wanted to give it a star, you had to drop your star fr om some other entry.

This was great in that it accentuated the positive but made people focus and think about relative merit.

The problem was that while the reusable star system was very good at identifying the top tier of entries, there were many second tier entries which didn't get the loving they deserved.

My second version included a "Green Dot". Everyone received a lim ited number of those, but quite a bit more (about twice as many Green Dots as you had stars total, if I remember). The two additional rules were that (a) you could only use them on entries which had fewer stars than some threshold and (b) your dot would automatically be removed and returned to you if the entry every crossed the threshold.

It was complicated to code, but easy to understand both by people "voting" and by people searching the library and I was very happy with how it worked.

The numbers of gold, silver, and bronze stars and the number of dots is proportional to the number of entries under consideration. I also eventually added a feature so that as the library grew, everyone registered occasionally received additional voting tokens.

Once it was all set up, it required almost no editorial management but provided a high quality mechanism for identifying the best submissions.

How about filters to remove certain makers from the view in the lists? I would really love a filter to remove version 1 and version 2 filters from view and concentrate on version 3 at least. I've seen about 99% of version 2 and they just clutter my scans.

I think you can be critical of filters to the maker as long as you provide something of WHY you are being critical. I never mind someone being critical as long as it's constructive.

I merely meant my ability to filter out filters, not everyone's view. I would never presume to restrict anyone else from viewing anything!

Guys... Me thinks you are wasting your time discussing any of this... FF is not going to implement any quality control because they want the numbers as a selling point...and certainly not going to let the community decide what filters get submitted... They are going to get you guys to create a quality control filter guideline for them though...