The problem with creating clips out of patterns instead of tracks, you get an enormous amount of tracks as each pattern can have different tracks right? I wonder how the play-engine deals with that… Our ideas are almost similar except I want to make the clips based on tracks (like any other DAW).
The “measures” still exists as visual bars within the arranger view where you can even choose to highlight an alternate X amount of lines, like how patterns are alternated in color now.
I think the best choice would be to provide an additional horizontal timeline view of the project in traditional arranger style, in addition to the pattern editor/matrix. I think the pattern matrix was a great addition to the current setup, but I don’t think it provided a solution to the underlying issue, which was what Knetter tried to solve as well, but I don’t think his solution again is again only a slight improvement. I think the arranger independent from the pattern structures would be the best choice; any change you make there to any point of the song wouldn’t be duplicated somewhere else but instead that pattern would become a new one. I think this would satisfy everyone since you wouldn’t need to use it and the current system wouldn’t be tampered with. It would also make easily possible to display (some) samples as waveforms 
(I hate the idea of showing them withing the pattern editor).
So, what are we talking about then? Patterns that we don’t call patterns?
Just another way of representing and arranging patterns, which allows you to escape the having patterns the same length on all tracks issue.
I think this would work well if you could toggle between a the current matrix arranger and this more other DAW like arranger.
One problem I run into with the current set up (I am not a hardcore tracker veteran) is what If I want to repeat something on one track every three beats, every 4 beats on another and every 6 on another. This can be very time consuming to realise with the current pattern arranger.
On the other hand, the current matrix is really fast for scetching up overall structure.
So why not this: The ability to toggle between the two.
But where pattern lenghts can be set for individual tracks, and in the matrix view you could choose to show the independent pattern blocks OR impose the block structure of a track of your choice onto the rest of the patterns, allowing you to arrange the song in chunks of certain time signature etc.
I’ll put up a picture…
There is a technique in music theory called suspension. Basically your extending the length of one note of a chord, and blending it to the next chord.
Chords are not the only application of course, but I hope you get the idea.
Similar sentiments of this thread were shared at the Pattern Matrix Rethink thread: post #25
If I’ve understood BotB’s description properly,
the issue occurs when he wants to shift a pattern so that it overlaps from one block
unto the next block, so this pattern is in two blocks.
I guess one way to avoid this knot sort of speak,
is to max out your pattern legnth in the hopes that it is enough, and in most cases it will be enough,
to write your musical phrase with this type of compositional technique or edit (suspension, overlap, etc.) in mind.
Just loop the pattern your in with the break pattern command ZB00.
There is the Rotate Pattern tool, perhaps a Shift Pattern would help ?
… i have no words. if you’re incapable of making a repetitive pattern without it being boring, then do not come here and demand bore’o’meters and other such filth. it’s not even funny, not by a long shot.
It’s also not like you’re coerced into using the same pattern over and over again to begin with, it’s just as easy to clone patterns as to create aliases, I almost never use the same pattern more than once. If you have 8 tracks playing a pattern and you want to variate one or two of them, what’s faster, cloning the pattern once and making the edits, or repainting each clip, making those two unique, and then editing? Clip based workflow introduces more steps to get the same result
You hit the nail on the head. Multi pattern sections are a pain to work with. Often you (I at least) want to shift a melody by just an X amount of lines: I have to copy the lines from the bottom, insert X lines on the next pattern, paste, damn now my pattern data is shifted out of the pattern, undo, paste in a temp pattern first, go back insert X lines, go back to temp pattern, select, cut, go back to pattern paste, repeat 3x…!! 
In an arranger it will be, click and hold the clip, move mouse, done! 
I like the idea of overlapping patterns of different lengths for visualizing poly rhythms, but I also think it’d be helpful when you want certain items to repeat. For example, if you wanted a kick drum every 16 steps you you could make one pattern that length and drag down in the matrix view to whatever total length you wanted. If you changed you mind later it would be easy to edit that one element.
I think the problem with using tracks instead of pattens is that you can only have one automation control, so if you wanted to do something with chords it’d be just as much work to keep all of those tracks organized and copying and pasting them together.
In the meantime, what about some copy and paste options for the matrix editor similar to the step size control in the pattern editor? It would control how many times asomething was copy and pasted and the empty steps between each copy. So for example if you have a 4 step pattern and you drag it down, move or copy it to another track within a 256 step pattern with “fill” it’ll paste every 4 bars until it reaches the end, but if you have it set to 6 it’ll have the 4 bars, move 6 empty bars and paste again until it reaches the end and so on.
Today I bumped against another very annoying issue that is indirectly caused by “The Great Wall of Patterns”.
It’s the automation editor. Renoise devs made a very cool line tool to edit across multiple patterns. However, since automation is mostly always multi-pattern it is impossible to make a cross pattern gradient of a value other then linear. Trying will be futile since you end up selecting the other pattern instead of changing the automation value. Also because the line tool is NOT creating a real automation line (it simple sets the value in the begin and end of each pattern it spans) you need to fix the begin value of each subsequent pattern.
If we break out of the pattern paradigm, we could get an automation lane like the instrument envelopes, otherwise we’re bound forever to linear value automations I’m afraid! 
Having the automation tied to a pattern allows you to create, “repetitive themes,” of automation… Patterns are everywhere imo… not just music… its literally “in our dna.” the beating of our hearts, the maze of our nervous systems… all patterns…
Music that repeats can be… is extremely exciting. that’s why… most music repeats itself… and forget modern day tunes… go listen to great classical composers of yesterday… patterns are everywhere in music, math, life…
I was thinking that too, sometimes it is preferable to edit the automation not being chopped up to the pattern length. I find it fairly common for a note sequence to repeat itself over several patterns while, for example, the filter automation changes constantly without a repeating. Think of dub step wobble, for example.
Back to the original post, leaving patterns out is not a good idea, it is simply a unit of time division: we have a beat, we have a bar, and we have a pattern. If we drop patterns out we’ll have a pulp of bars with nothing to distinguish them between each other. A big no-no.
Look at the pattern matrix and then honestly do you look at the pattern as a box or at the little lines, notes, that are in there? I guess the lines. Now imagine we take the block drawing out and make it just one big-zoomable-set of lines. You can put time code markers here and there or arrange a repetition in a clip. That’s how I would like to see the UI without patterns. I started this thread to highlight the structural problems/limits with the pattern approach. With the current automation bird-eye-view you see some real quirks and I reckon we hit some kind of wall for being innovative in the sequencing UI.
Don’t get me wrong, but the idea you have of patterns in the more logical sense I fully share. But 
. I don’t imagine them as a fixed set of tracks over a certain period. Patterns can occur on different levels in the music, even with different lengths. Therefore they need to become “scoped to” separate tracks: hey, this is the clips idea. In the end a clip is also some kind of pattern but on a per-track basis. Much more flexible. And if you want you can still group all your clips and cut them off at the same point to have the original pattern back (hence I bet you’ll never do that).
Even though we’re using computers with tons of RAM and processing power, there’s still a little niche, where this feature would be extremely handy. In size-restricted competitions (4 kilobytes and 64 kilobytes intros) where softsynth, note-data and various algorithms for procedurally generating audio and visuals are packed into small executable files. When doing music for such productions, having the beforementioned feature would be fantastic! Complexity/variation has to be balanced as it’s always at the cost of size. Repitions packs better than variation, obviously 
An example of what is possible (if accepting enough repetition)
That’s ofcourse a video capture. The executable requires a rather high-end graphics card
There is no structural problems/limits about patterns. You are simply trying to cram a FT/Reason/Cubase/etc sequencer paradigm into a tracker.
However, track automation does not work very well with the current approach, but I don’t think clip oriented one would be the only usable solution.
The Buzz Sequence Editor does exactly what you want - if only they would implement it in Renoise. It looks like we are going to be offered anything BUT the most simple solution (the Buzz method) and will never get a working solution.
Am I right in thinking that you’re an American?
No, he is from Netherlands.