It is a universal rule of all engineering practice (I am also an EE) that a system incapable of producing a required output, such as a refrigerator that will not cool, is seriously deficient. In the software trade we call these deficiencies bugs. In the case of the orchestral staff attribute, the standards of music engraving demand that all and only those staves so marked should have an orchestral bracket, and that the tops and bottoms of each such bracket in the system must have little circular finials. NoteWorthy puts a bar across the entire system if any stave is marked orchestral, and the finials appear only on the top or bottom staff of a system. The designer of a system that is to interface with the outside world in which standards already exist does not have a free choice of the way he wants to handle such items. I provide four examples from the 138 scores I have made Worthy of Note in the last three years, and will testify under oath that I have seen no counterexamples.
This one wart on its nose makes NoteWorthy a joke as a serious printing program, for all the wonderful things that have happened with respect to slurs, hairpins and stem controls. It is all the more ironic, then, that after finally reviewing the entire example directory in detail, I find that NoteWorthy can produce a virtuoso performance, at least if the performer it emulates is a precision monger like Glenn Gould. For God's sake, get this fixed, issue NW2, and start working on NW3, which should print, understand and play all the standard abbreviations for tremolos, repeats, arpeggios, trills, mordents, turns, appoggiaturas, acciaccaturas, glissandos and whatever else one can find in ten minutes of googoling, as well as provide an alternative to dragging over items to select them, and the ability to select and modify a note in a chord, as well as a fully floating C-cleff (see the Chorus of Peers example) as well as countless other rarely needed but essential when required bits of feature and notation.
As promised, here is described the horror of using NoteWorthy to engrave a score which needed to be presentable enough to be used for more than the composer/arranger's scribbling.
Toward the end of 2006, I was given the job of preparing NoteWorthy scores for the eighteen works to be sung in May 2007 by the choral society of which I am Librarian and Baritone Voice Leader. Mostly this was reasonably straightforward, as Sharp Eye is rather good at scanning good quality scores published in the last fifty or sixty years. Three of the works, however, were in manuscript, and the one here described was the worst of the lot. As a quick glance at the scanned page attached as Land Sighting001.gif will reveal, it would be hopeless to expect anyone but Robert Wilson, our former accompanist and creator of this scribbled mess, to play or sing from these pages.
Believe it or not, Sharp Eye actually found all the bar lines and nearly a third of the notes, which provided a sufficient base from which to complete the score using it's truly marvelous proofing facility, which I believe is faster than anything but a perfect performance on a MIDI keyboard, which I have, but not sufficient skill to use. I then passed the output of Sharp Eye through a midi file to NoteWorthy, converted the unholy messes of tied notes that represent triplets (how about detecting these, at least most of the time, automatically?), expanded the abbreviations for the numerous extended trills and replaced the symbols for repeated figures with copies of the figures in question.
At this point, I had eight staves, one for each choral division and two for each hand of the accompaniment, which is probably the minimum for any notation processor which actually tries to understand the music, as it must to play the score, and this was suitable as a home study aid where the voices really wanted to be separate and the keyboard parts could be suppressed. I then went through the rather subtle drudgery of hiding redundant notation, and adjusting notes that would otherwise improperly overlap when the staves were layered, to make a score suitable for printing.
Here I encountered the "Orchestral Staff Property" bug. I'm sure everyone reading this knows what it is, so I won't belabor the point except to note that it is listed as fixed in NoteWorthy 2 and has a line in the "What's New" section of the help file. They lie. The behavior of NW2 and NW1 are identical. As it would be absolutely unthinkable to either omit the orchestral brackets or present the results NoteWorthy produces in a score to be displayed in public, I drew on my expertise in document restoration and forgery to correct the problem graphically, using the two NW2 files attached, printed and scanned, and as it turned out, I needed to do this twice.
I blithely prepared and had printed a nine page score for the piece, and distributed it with the rest of the music at our first rehearsal of the spring term in February 2007. The print proved too small for most of our membership to read, so I prepared a score in what was judged by a committee to be the smallest easily readable point size, and this is the twenty-two page version I would love to attach as a Microsoft Word file. In several places, I could get only three measures per page, (one such measure is attached as Land Sighting 3 008.gif) and so the accompanist who replaced the virtuostic Mr. Wilson got a page turner for the first and last time in my tenure with the club, and still struggled with the piano part, which would in fact be the most interesting feature of this piece, if played really well.
Conclusions
The best way of solving this problem would be for NoteWothy to understand things like the beamed open headed notes indicating a trill and the big sort of percent sign that indicates a repeated figure. While you are at it, putting in notation for arpeggios, turns and other ornaments, also to be understood by the notation to midi-event processor, as simple grace notes are now, would be a good idea. In the end, the "Moonlight" example file should play as it does now, but with no hidden staves.
Failing this, nearly all numbers in NoteWorthy could be made signed floating point, including the "extra" note space attribute. The staff position attribute is already signed, and I understand from discussion elsewhere in this forum that the internal representation of things like stem lengths are already non integral, so this massive but straightforward overhaul, mostly but not completely confined to the header files, should be doable. True, a meaning and print representation for a tempo indication of "-33.333333" must be found, and notes probably shouldn't be allowed to migrate to the left of a note preceding them (though overlap really should be considered ok), but I have always found that being maximally canonical in language definition is best for both the user and implementer. (My day job, before I retired, was in software development.)
Finally, to those who have reacted to my previous recitals of obvious bugs with drastic consequences and minimal effort to fix by saying "what do you expect from a beta level product", I can only ask how many multiple years are legitimate for a so called beta test? Or is NW2, for all of it that has been done, really just vaporware? I hope not. NoteWorthy's unique ability as a superior sound-attached-to-notation generator is something I and my Club depend on, and while we could survive on NW1 indefinitely, I really would like to be able to print a score and not be embarrassed by it's appearance.
NoteWorthy simply and flatly has a bug in it's handling of the "orchestral" staff property. It puts a heavy accent from the top to the bottom of the system if any staff has the orchestral property and a finial only on the top or bottom staff and then only if that staff is flagged orchestral.
The correct behavior is to put a heavy accent on and only on each staff that is orchestral, and put a top finial at the join between an accented staff and an unaccented or nonexistant staff above and a bottom finial at the join between an accented staff and an unaccented or nonexistant staff below. The only way I have found to get around this problem is to print pages, scan them, and paste in the accents and finials graphically.
This one problem alone makes NoteWorthy nearly useless as a serious printing program, despite all the added prettyness from version 1, and is listed as fixed in version 2. So where is the fix?
