TromboneChat

Great article on AI 'teaching itself':

"When AI's Output Is a Threat to Itself"

https://www.nytimes.com/interactive/202 ... Position=1

"When A.I.'s Output Is a Threat to Itself"

NY Time, August 26

https://www.nytimes.com/interactive/202 ... Position=1

I just want to take the opportunity to say this on trombonechat: I love R. There said it. Never thought that it would come up on this forum.

Thanks for this thread. I don’t have a current license for Finale annd would have missed this news. I will have to dig up the files I have and see if my brother-in-law can convert them to xml for me. He’s a music publisher and I wonder what he will do.

BGuttman wrote: ↑Thu Aug 29, 2024 11:38 am There have been changes in music software all along the way. I used to use something called Encore (by Passport Software) back in its heyday. Encore is about as current as Fortran (another language I know) and Cobol. Things you learn about in the History classes. Encore was OK in its day, but Finale and Sibelius put it to shame.

There is a point where trying to patch an old system to keep it running is more pain than gain. Windows 95 and NT supplanted the old MS-DOS/Windows combination in the mid 1990s because trying to keep updating the old system became unsupportable. Finale has been through a lot of upgrades and I suspect it's getting harder and harder to figure out what is going on what with the patches and alterations. A clean start is needed.

I worked for Passport Designs during its last year or so, before it sold all its products to another company. The story was that, once they had established market leadership in music notation software with Encore, they turned all their attention to developing a multimedia authoring product. During the time they stopped paying attention to Encore, Finale caught up and surpassed it, and Sibelius was right around the corner.

All Passport wanted to do at that point, and the reason I was there, was to develop a 16-bit Windows version in order to make their product line more attractive to a potential buyer. It was a fun job while it lasted. I wrote some pretty fair music in the process of testing Encore.

So farewell Encore and Finale. We will soon say farewell to Sibelius, and who knows how long their successors will last. Don't get too dependent on any of them.

The sort of programs AI will replace are all the ancilliary IT stuff that every company builds up over time. All those spreadsheets created by people to help keep track of their own jobs, that give IT departments sleepless nights because nobody knows how they work, have no automated tests and are likely hiding nasty corner case bugs left right and centre, will instead be created by AI. The fun is they'll continue to keep IT awake at night for all the same reasons, but there'll likely be even more of them with a lower bar to entry. That one person in the department who's revered for knowing how to do pivot tables will instead be revered for their AI prompts.

This talk of "all you need to give it is a complete spec" will never happen. Our industry has tried again and again for decades to "just write a complete spec" and transform it into a working program and it's failed every time. I've lost count of the different forms specs have taken with this explicit aim, and they've all failed. Outside a few extreme niches where a handful of programs are specified in formal mathematical logic and then programatically converted into running code (and I really do mean a few - it must be fewer than 10 programs written like this on the planet), it just won't happen. I think of it like the recurring attempts to "just write programs in plain english". It seems every generation needs to learn this lesson for themselves - that natural language just isn't precise enough to describe how complex processes work in every scenario without becoming even harder to read and more verbose than a programming language.

mgladdish wrote: ↑Fri Aug 30, 2024 12:47 am The sort of programs AI will replace are all the ancilliary IT stuff that every company builds up over time. ...

This talk of "all you need to give it is a complete spec" will never happen. Our industry has tried again and again for decades to "just write a complete spec" and transform it into a working program and it's failed every time. I've lost count of the different forms specs have taken with this explicit aim, and they've all failed.

Well, I do understand this perspective, and the history on which it's based. But I think you're behind the times in terms of the current state of AI, the significant change in methodology in natural language understanding (since the 1990s), what I'm inclined to think of as advances in intelligent "inferential" and "modelling" technology, and the degree to which "AI systems" are being deployed in various industries at this time -- including, for example, medical contexts for diagnosis, reading and interpretation of radiology tests, etc., and business applications (where it acts as an informed and interactive "consultant" in real time -- my daughter, working at one of the world's largest risk analysis companies, uses this on a daily basis) This really isn't at all about having AI "write programs" from some higher level program spec you provide to it.

But time will tell.

I'm curious to hear how other Finale uses are planning on transitioning now.

I've been using Finale since 1989 (I believe it was Finale 2.0). Over the years I've gotten pretty good at using it and know exactly how to get things entered and looking the way out want pretty quickly. The most frustrating thing for me won't be transferring my Finale files over to other software, it will be the learning curve on the new software.

For the most part, I already have my Finale music printed out as PDFs already and most of that stuff I'm unlikely to change because it's already completed. I plan to get most of it imported over anyway, but I'll probably take the opportunity to just get rid of some files that I'll likely never need again.

I went ahead and bought Dorico. I've only used it once so far, just to go through some of the tutorials. I think that the major differences won't be such a big deal once I learn how to do things, it's the little things that will mess me up. For example, on Finale pressing 5 indicated a quarter note. On Dorico a quarter note is 6. Entering notes using Speedy Entry on Finale I don't really need to think so much about what keys to press for what rhythms, so having to adjust that is going to take me some time.

Yesterday I needed to take a standard tune (Mercy, Mercy, Mercy) and put together a transposed part for a student who doesn't have the necessary fake book. I thought about doing it on Dorico, but chickened out because I knew I could crank it out in Finale very quickly. At some point I will need to bite the bullet and stop using Finale.

Dave

Wilktone wrote: ↑Thu Sep 05, 2024 8:34 am For example, on Finale pressing 5 indicated a quarter note. On Dorico a quarter note is 6.

I know nothing about any of this software, but isn't this configurable in Dorico? I'd like to say that I'd be astonished if it isn't, but too often the UI "designers" of applications like this seem to leave out a bunch of fairly obvious UI design. In fact -- even as just encouragement for Finale users to convert over -- I'd expect them to provide a simple export/import mechanism to make such a transition as painless as possible. Do they really not have anything like that?

I dunno yet, I've only started to scratch the surface of Dorico.

Finale had "metatools" or whatever they ended up calling it, so that I could set certain keys to certain articulations or dynamics and such. I don't believe that Finale allowed you to change the Speedy Note Entry keys, but to be honest I never looked for that.

My 2 cents as a software dev who has to hop between all sorts of different tools and software all the time:

Bite the bullet and learn the new tool as-is. Even if Dorico does allow you to change all the key mappings to the same as Finale, you'll still be swimming against the tide. All the online help will be in terms of the default mappings and there'll still inevitably be some mismatch somewhere. It takes a lot less time than you think for new keyboard shortcuts to become ingrained. So it's better to just stick with the defaults.

I'd also not be so quick to slag off the teams that built any of these tools. What may be obvious UI design to one person can be an unusable nightmare to another. If you're interested in the reality of it all there's a brilliant pair of videos by a software designer looking in-depth at the UX of both Sibelius and Dorico. I'd strongly recommend a watch.





I'd also urge a bit of restraint when wailing about imports/exports between Sibelius, Finale and Dorico etc. Each tool has its own strengths and weaknesses so there will never be a complete 100% compatibility between them. Take Microsoft Word and Apple Pages for example - Apple have more money than god and still can't get MS Word docs to import and retain 100% of their formatting in all cases. And that's a usecase that has literally billions of pounds payoff if they could get it right. So it's not realistic to expect a relatively tiny industry like music notation to succeed where the unimaginable megabucks of Microsoft and Apple have failed. They're about where I'd expect them to be - imports mostly kinda work, but expect to do a bunch of tidying up around the edges. And it's definitely better than starting again from scratch even with printed parts as reference.

mgladdish wrote: ↑Thu Sep 05, 2024 9:20 am Bite the bullet and learn the new tool as-is. Even if Dorico does allow you to change all the key mappings to the same as Finale, you'll still be swimming against the tide. All the online help will be in terms of the default mappings and there'll still inevitably be some mismatch somewhere. It takes a lot less time than you think for new keyboard shortcuts to become ingrained. So it's better to just stick with the defaults.

In general I agree with this -- if for no other reason than often a new product simply does things better and with higher usability than an older one. Just differently. I think this is called "improvement" or "progress". Nonetheless, various users will take time to adapt to it, and that will require some work.

I once created an enterprise-wide "file comparator" product for graphically displaying differences in text files (specifically to be used for C/C++ code files for developers, but not restricted to that). The old version had been clumsily done and did its comparison based on an algorithm of detecting when it saw a difference and then using "look ahead" to determine when to "resync" the matching lines in the file. I replaced it with a much better/faster/more accurate approach (a well-known comparator algorithm) that produced a maximal match and made comparison and editing much more obvious -- plus a much better graphical UI. After it was deployed, a sub-group of users would call me and tell me that it wasn't working because it was "getting out of sync" with the files and that it should be "re-syncing earlier" -- completely ignoring the final result which was in fact more accurate, much faster, less convoluted, and much easier to use. Being developers themselves, they had become wedded to the implementation rather than to goal-oriented performance. But people tend to like what they're used to -- I know I do. (When last I visited my old friends at that company, they said that they were still using that file comparator application I had written -- about 20 years earlier -- even though by that time there were much better tools and environments to supplant it.)

I do see quite a number of documents and YouTube Videos on moving from Finale to Dorico. And while it's annoying to "suddenly" have to spend your time learning a new approach (which does take time away from doing the work that the tool is meant to facilitate), it is almost always better to adopt/embrace the new (actually improved, one hopes) model rather than cling to the old. Hey, I just switched off MS Windows and onto a Chromebook!

Dumb question. Can the current version of Finale continue to be used for old Finale files even if it doesn't get updated?

I understand that if you want to do latest and greatest new work you will have to migrate away from Finale, and your choice may be dictated by the choices of the publishers you work with, but will you have to go through with he heartache of migrating hundreds (perhaps) of files as well?

BGuttman wrote: ↑Thu Sep 05, 2024 12:45 pm Dumb question. Can the current version of Finale continue to be used for old Finale files even if it doesn't get updated?

I understand that if you want to do latest and greatest new work you will have to migrate away from Finale, and your choice may be dictated by the choices of the publishers you work with, but will you have to go through with he heartache of migrating hundreds (perhaps) of files as well?

On Windows, the answer is probably yes. Especially as they've agreed to keep the machine running for licensing. For Mac, the answer is less obvious. They've made some very bad breaking changes to the platform on an OS level and that has affected older versions of Sibelius as well as Finale. It's no guarantee that you'll be able to keep running Finale indefinitely unless you are running on an old version of Mac that no longer receives updates and then you'll of course be at the whims of the hardware not failing.

An experimental posting:

I started late 1980's (and had my own engraving company until around 2011) with HB Music Engraver and lost everything a couple of years later, when HB Music Engraver folded and an OS upgrade prevented the program from even starting ever again.

When Finale was just trying to get professional recognition, SCORE was THE professional program. Finale was considered to be cute, but output quality wasn‘t worth the effort. This POV led by Henle and at the time, and what Henle said, counted. Well, when Leland Smith died in 2013, orphaning SCORE, guess where Henle moved to…not Encore, which also left a loyal following high and dry, no .. Henle bit the bullet and didn‘t move to upstart Dorico, it went to Finale. I wonder …

Meantime, I’ve always kept full PDFs of my work since mid-to-late 1990‘s - long story. I examined my records and, lo and behold, I‘d paid around $149 for the last couple of Finale updates. So I cleaned out a 2012 MacBook Pro (keeping that deep in reserve) and my beloved 2015 MacBook Air, paid my money and bought the Dorico cross grade to gain access to the last Finale version, DL’d Finale v27 and installed on the 2015 Air and an M2 15" MacBookAir ... as opposed to statements here, neither v.25 nor v.27 have problems running on the new Mx processors nor old Intel iron - see below. All went well until I went to authorise and while the sale with the new serial number was recorded in my MakeMusic account, the authorisation didn’t work. Same old non-response from MakeMusic, when in an act of desperation I erased the installations and did them again - and this time the authorization magically worked. And one wonders why Finale went belly up. I’ve also been reading about the difficulties installing Dorico and unresponsiveness of Steinberg - life goes on as normal. Now with the most current version of Finale installed on two good machines (along side v25, which still works fine), one of which that cannot be and won’t need to be upgraded Finale machine, I have breathing room to investigate viable alternatives or just plain play my bass trombone and play with my dog.

Each of the remaining big guns have camps of proponents and hordes of detractors, most spreading misinformation, downright lies and "half truths" about their favorite product and about the competitors'. Before I definitively go to another company (I've DL'ed versions of Sibelius for which I was a beta tester up through version 1.4, Dorico and MuseScore), I have some files that display what I need to do, I learned to do quickly and well in Finale, and will see how the programs handle them, respective communities, as well as how/if official tech support handles inquiries.

ghmerrill wrote: ↑Thu Sep 05, 2024 8:46 am

Wilktone wrote: ↑Thu Sep 05, 2024 8:34 am For example, on Finale pressing 5 indicated a quarter note. On Dorico a quarter note is 6.

I know nothing about any of this software, but isn't this configurable in Dorico?

I've done a bit more research and it is customizable in Dorico.

mgladdish wrote: ↑Thu Sep 05, 2024 9:20 am Bite the bullet and learn the new tool as-is. Even if Dorico does allow you to change all the key mappings to the same as Finale, you'll still be swimming against the tide. All the online help will be in terms of the default mappings and there'll still inevitably be some mismatch somewhere. It takes a lot less time than you think for new keyboard shortcuts to become ingrained. So it's better to just stick with the defaults.

That is a good point, but I had already customized quite a bit using Finale and didn't really find it too difficult to work with online help when I needed it. But I think I'll work with Dorico for a bit as is before I start customizing key bindings and such. Once I get more comfortable then I'll start tweaking things.

I'm sure that I'll set up my own customizations soon. There are certain things about the way I work and certain things that I'll need to do more frequently than other users and setting up my own key bindings is an attractive idea.

Dave

Wilktone wrote: ↑Fri Sep 06, 2024 7:39 am
I'm sure that I'll set up my own customizations soon. There are certain things about the way I work and certain things that I'll need to do more frequently than other users and setting up my own key bindings is an attractive idea.

That is, after all, why they have designed that feature into it; and you shouldn't be reluctant to use it.

BGuttman wrote: ↑Thu Sep 05, 2024 12:45 pm Dumb question. Can the current version of Finale continue to be used for old Finale files even if it doesn't get updated?

Not a dumb question.

Yes, it can. The main concern is an OS upgrade breaking the software. I'm on a Mac, and as Matt mentioned, upgrades to the Mac OS has broken Finale in the past.

For now, I'm keeping Finale installed as well in case I need to access it.

Dave

Wilktone wrote: ↑Fri Sep 06, 2024 7:42 am The main concern is an OS upgrade breaking the software.

I think this is the biggest problem with the software products we use nowadays (well, aside from crappy design, often being blind to genuine user needs, and general lack of testing prior to release). In the ancient days (mostly prior to the dominance of the web and client/server use for virtually all applications of any consequence), you could install an application and it would run as long as your machine and OS did. Any application updates would pass you by, but leave you happily using the outdated version. Applications were pretty self-contained. You can still write such an application (if you're careful), but it won't be able to take advantage of the capabilities that users want in their 21st century applications. Nobody really wants to be running a line-oriented DOS (or ksh or bash, or ...) shell any longer.

So now the problem isn't just what most people think of as an "OS upgrade", but the dependence on dynamically linked libraries that the OS or the app may use. The change that breaks backward compatibility may not even reside in any direct OS upgrade, but in some library that the OS makes use of for a purpose only indirectly related to the application that's affected. Such is the price of the progress we're seeing in computer and software use.

This isn't anything new, the problem with Mac OS is they - particularly recently, but also historically - have done major breaking changes from OS version to OS version, which are released much more frequently than Windows major version updates. Windows 11 can still run ancient version of Finale and Sibelius (I've installed as early as Sibelius... 4(?) on Windows 11).

Its sad what seems to happen to all notation software these days. I use both Finale and Sibelius. Before I used Encore Passport. And in the end of 1980, maybe 1987? Atari Notator. I liked all of them. Today I have lots of old arrangements with these old software wich I cant reach anymore.

Maybe the survivor will be Musescore wich is free? There is no big company behind it. Anyway I get the Encore back soon on old diskets. Cant wait to see my old scores again.

Leif

Wilktone wrote: ↑Fri Sep 06, 2024 7:39 am

ghmerrill wrote: ↑Thu Sep 05, 2024 8:46 am

I know nothing about any of this software, but isn't this configurable in Dorico?

I've done a bit more research and it is customizable in Dorico.

mgladdish wrote: ↑Thu Sep 05, 2024 9:20 am Bite the bullet and learn the new tool as-is. Even if Dorico does allow you to change all the key mappings to the same as Finale, you'll still be swimming against the tide. All the online help will be in terms of the default mappings and there'll still inevitably be some mismatch somewhere. It takes a lot less time than you think for new keyboard shortcuts to become ingrained. So it's better to just stick with the defaults.

That is a good point, but I had already customized quite a bit using Finale and didn't really find it too difficult to work with online help when I needed it. But I think I'll work with Dorico for a bit as is before I start customizing key bindings and such. Once I get more comfortable then I'll start tweaking things.

I'm sure that I'll set up my own customizations soon. There are certain things about the way I work and certain things that I'll need to do more frequently than other users and setting up my own key bindings is an attractive idea.

Dave

Well, you would be a perfect candidate for buying a StreamDeck. It'll change your notation life!

I bought one several years ago, for Sibelius, and set it up with keystrokes, pages, input shortcuts, etc. etc..

Now that I've been using Dorico awhile, I kept most of the same shortcuts that I like, but Dorico offers so many more built-in shortcuts (things that were all 3rd party in Finale or Sibelius are built right into Dorico!)...that it's hard to imagine not owning a StreamDeck now.

For more information, check out Scoring Central and their Notation Central. They offer templates to get you started and much more, and then you can go crazy with your own customizations.

https://www.scoringnotes.com/meta/notat ... -released/

https://www.notationcentral.com/product ... ductivity/

When it was announced, that Finale would no longer be sold or supported and the upcoming Sequoia OS version would not be certified as compatible to Finale (funny that choice of terminology, NOT that Finale would not be compatible with Sequoia....), I bought for $200 a pristine 15" 2015 top-of-the-line MacBook Pro with a new screen and new battery.

Alternating between it and my daily driver, a 15" M2 MacBook Air, I've done a series of 6-12 Sonata sets of cello/continuo pieces. I started out doing them in Finale v27 (which was new for me and it's different than the v25 I had been using), Dorico and MuseScore 4.4.x.
right off the bat, Dorico and I didn't get along. Confusing, confounding, stifling, and that was just the program - the on-line community was even worse - their standpoint "Dorico's way or the highway". I gave up the fight with Dorico after about 10 days and not being able to complete a single movement of a single sonata.

MuseScore was different. Different in layout, different in "philosophy", different in logic, different is ease of learning. Just as complex as Finale, just as capable of custom configuration as Finale, maybe a little easier to do so. Some functions of MuseScore are simply better than Finale - like figured bass, which is a multiverse HORROR in Finale and a paradisal joy in MuseScore. But it does have its glitches.
Note input is at about 80% the speed in MuseScore that I can do in Finale, which is acceptable. And there are a couple of glitches here, like the input point for no apparent reason jumping an octave - there's a quick correction for this, so I got very quick and quickly recognising this aurally and correcting.
Then it came to me that I could take the strengths of both Finale and MuseScore and work out a workflow starting in one program and finishing in another. Using XML to communicate between the two. Easier said than done. It turns out, there are problems with XML in this regard. And in typical computer industry fashion, when dealing with the support gurus, it's always the fault of THE OTHER program. I don' wan' to lay blame, I want to now how the work-arounds work. There are none. You just have to know what breaks and how to fix it. For example:
1) Importing XML made in Finale - no matter which format - measure numbering, which is time consuming and fickle in Finale breaks when imported via XML into MuseScore. No big deal. Just go through each movement of a 12 sonata set - about 48 movements - and set a "section break" marker at the end of each movement, which automagically then starts the next measure on a new system with a new measure "1". If it's a pick-up, you have to tell MuseScore to ignore this pick-up measure in the measure numbering. You also have to check all mid-measure repeats and do the same for the 2nd half of the repeated measure.
2) Although all the note material survives the move via XML between Finale and MuseScore. Key signatures "act" wonky under certain circumstances after such an XML-trade. For example, if you want to transpose a movement down a step from C to B-flat, and either use the key-signature tool OR the transpose function, only the notes get transposed, the key signature stays the same. So you get a key signature of C and the notes in B-flat with "accidentals". Fix: Go through the entire collection, movement for moment as in point 1 above, and re-set all key signatures, even when they are showing correctly. So Even if C major is showing correctly, you have to re-apply the key-signature of C maj. THEN, after doing this throughout the entire set, if you transpose, it works correctly.

There's more, much more, but still, there is a work-flow that works for me. If a piece moves for the most part step-wise, Finale is the best choice for note input. I can also analyse it BEFORE adding editorial dynamics (font problems between the two programs) and do any transposing within Finale. THEN save as XML and open this file in MuseScore. Within MuseScore add editorial changes and do page formating. Page formating is "better" in MuseScore with few exceptions. If the piece includes large intervalic jumps and lots of them, using MuseScore octave-jump-beg as a feature, it can be actually easier and quicker to notate such a piece in MuseScore and skip Finale altogether.

Back to hardware: ignoring computer start-up times, both the "modern" M2 and the vintage "Pristine 2015 computers have no problems handling Finale or MuseScore. Or LibreOffice or YouTube for that matter. I used Open Core Legacy Patcher to upgrade gthe vintage 2015 boxes (two now) to the same OS version I've got on the M2 and it works just as well and just as stable. I will take the M2 to Sequoia after New Year, meaning after a couple of newer bug fixes. I will keep the "vintage" macs on Sonoma 14.7.1, because they work perfectly well, after fairly current in terms of security (and I don't travel in the wild with them, so they're always behind my secure firewall at home) and Apple isn't going to release anything new for them that will, when installed, break a running Finale system.

These sets of Sonatas and some other stuff before them, can be found here:
https://5d832781b3df5.site123.me/free-trombone-music

The new pieces done with Finale, MuseScore or a mixture of the two:
12 Cello Sonatas, 1-6, Pierre Hyacinthe Azaïs (1741 – 1796)
12 Cello Sonatas, 7-12, Pierre Hyacinthe Azaïs (1741 – 1796)

Azaïs_12 Cello Sonatas 7 - 12 - Bass Trombone, SCORE.pdf

6 Sonatas for Bassoon or Violoncello, John Ernest Galliard (1687 – 1749)

6 Sonatas for Violoncello, Giraud, François-Joseph (? – 1788)

6 Sonatas for a Violoncello, Op. 1, Willem de Fesch (1687 – 1761)

6 Sonatas for a Violoncello, Op. 8, Willem de Fesch (1687 – 1761)

6 Sonatas for a Violoncello, Op. 13, Willem de Fesch (1687 – 1761)

de Fesch, Six Sonatas for a Violoncello, Op. 8 - Bass Trombone,SCORE.pdf

5 Sonates suivies d'un concerto, Op.26 (ca. 1729), Joseph Bodin de Boismortier (1689 – 1755)

Six Solos for a Violoncello and a Bass, Op. 11 - Gian Battista Cirri (1724 - 11. Juni 1808)

Another "problem" with MuseScore/Finale is figured bass. Besides figured bass in Finale being a nightmare, it doesn't understand MuseScore's iteration of figured bass when a MuseScore file with figured bass is saved as XML and imported into Finale. No figured bass comes through. If youo take the same XML file, just exported from MuseScore and re-import it into MuseScore, the figured bass is there again. so MuseScore DOES actually understand figured bass when exporting XML. Finale simply doesn't understand MuseScore implementation in XML of it's own figured bass.
And a further "problem". I've yet to find a way to create dynamics using either the Finale native font or the MuseScore native font with a bracket around it, that signals, that this dynamic is editorial. Even if I do so in a font found on both programs for the brackets, they don't come over through the XML process in the other program. Not even when re-importing into the same program that just created it.