Comments in [coll]

    Jun 01 2011 | 5:05 pm
    Dear list,
    is there a way to put comments in a text file that is used by a [coll]? I'm having a patch where performance parameters are stored in huge colls (a quite frequent scenario, I guess) and it would be very, very helpful if I could add some small comments here and there that would be ignored by the [coll] operations and outlets.
    Thanks, Ádám

    • Jun 01 2011 | 5:52 pm
      something like this (you may ignore the "s" in the unpack)?
    • Jun 01 2011 | 6:43 pm
      Thanks for the reply. However, I'm looking for a solution where I don't get the comments out of the [coll] outlets at all, for two reasons:
      1) Independently of my actual problem, I'd like to know whether this feature exists, as I would find it a quite useful feature for anyone that has to deal with [coll]s.
      2) In my particular case, since my [coll]s don't have a fixed number of columns, and some of my algorithms processing the outputs will act depending the number of items contained in a row, it would take much more time to re-write the whole patch in a "comment-aware" way than to take notes in a separate text file (which I'm currently doing).
    • Jun 01 2011 | 7:10 pm
      I've used parallel colls for this in the past, which is a PITA, but if you're managing the data yourself, and not relying on the coll editor to enter data, is not completely horrible.
      I agree that comments in colls would be a good feature.
    • Jun 01 2011 | 7:30 pm
      I show a way to do this, albeit by not using the coll file format at all, then parsing and loading into a coll, here:
      best, Zachary
    • Jun 04 2011 | 1:19 am
      Wow! This tool, although in my current case I can't benefit much of it, is really cool! I'll keep the link for future reference :-)
      I don't know if there is an 'official feature-suggesting' forum (I guess there isn't), but regarding comments in [coll]s, it shouldn't be hard to implement. Since the data in the [coll] must be ended by a semicolon, and the identifier of each row is a single atom (either a symbol, a long or a float -- so, no lists allowed anyway), I don't see why the content after a semicolon but before the next identifier why couldn't be treated as comment/whitespace... (Now the editor simply gives an error if I try to do that, so the internal code does even recognize that something's at the wrong place -- if instead of giving an error it would simply be ignored, then we could use that place to put comments...)
    • Jun 04 2011 | 3:35 am
      Could you not code something around [zl ecils 1] to always remove the last item at each [coll] index? It might take a bit of work for a few of the commands but for normal store and recall it should be fairly easy to implement.
    • Jun 04 2011 | 7:54 am
      Thanks Siska, glad you like that tool. I'm using it all the time.
      Here's I think what you're asking for. You can just add comments to any coll file after the end of line semicolon, or in its own line. The patch parses out the comments and then loads into a coll. A nice clean and reusable abstraction can be made from this with a little time...
      best, Zachary
    • Jun 06 2011 | 11:18 pm
      yep, this is a nice and generic solution.
      I'm wondering if the Cycling '74 developers could add this simple behaviour to [coll] itself in a later release (maybe Max 6?), though. Theoretically I don't see why this couldn't be implemented, as it's quite easy and it doesn't cause any backwards-compatibility problem...