I have a question and I don't want anyone biting my head off and suggesting I'm being critical of Hash or the wonderful work Stefan and the other programmers do. I really do love the program and am continually astonished by what it can do for the low price point.

So here I am, almost a year since I bought my CD (I think I have about a month left) and I'm curious about the version thing.

(Here is the part that sounds critical but honestly it is just a question I would like an answer to. Also my question assumes that version 16 will be out early next year... that may not be the case, they may keep on with iterations of 15 for some time. I hope you can see that the question is still valid without me altering it to explain what I mean should 15 L or something be the last version I can access with my CD purchase)

In the year since purchasing there have actually been a surprising number of updates to the program. 15i was just released? That's the ninth iteration of Animation: Master 15? It seems that while each version has some new features and bug fixes, almost every time a version is released a crop of new issues arise and the guys at Hash start working on the next version and including fixes to those bugs there.

So I'm wondering a couple of things. First I'm wondering if we're ever going to get to a "safe" version of the software. By safe I mean a version where the developers say "we are adding no more features at this point, all reported bugs are fixed, this is the most stable version of this iteration of the program. This is the point we will start at when we start making changes for version 16."

Second, if this is not the case, if there is no "safe" version of A:M 15 planned then does that mean there will be no support for 15 after 16 is released and I will just have to find the one version among the many that is most stable on my machine and live with whatever flaws it has?

And lastly is the only solution to this issue to buy 16 when it becomes available and keep updating in hopes of finding that one perfect version of the software that has no bugs and works perfectly on my machine?

Like I said, I'm just looking for an answer, not trying to give anyone a hard time. But I need to know if, when it comes time to buy the subscription if I will be able to choose to buy it because I want some new feature or if I will be forced to buy it because I don't want to live with bugs and updating is the only way past them. I don't exactly tax the programs abilities. I'm just not that learned yet. But things like the random closings I'm getting in 15g are not something I'm going to happy living with. Obviously I can now upgrade to 15i and hopefully that issue is resolved. I guess I'm just trying to figure out if I need to put aside (very precious right now) funds for buying a subscription in Late January or if I will have a stable version by then that will allow me to put off spending the money for awhile.

Hope this doesn't anger anyone. Just trying to plan for the future.

For your current definition of 'safe' the answer is no.
There is a point in each version where a plateau is reached; where the trade off of benefits going forward outweigh the safety and security of staying where we are.

So we can move on to your next case...



If you look at the updates reported for each release you'll see that fixes are often found for previous version bugs.
The problem is whether to spend time going backward to fix previous versions or concentrate on fixing things that are currently 'on'.
This is where the subscription model comes in.

Keeping our subscription current allows us to keep pace without expecting development to go backward.
Hash Inc doesn't have the resources to support previous versions but can support users via the current release.
There are simply not enough development dollars. I'm sure Hash Inc would support these older versions if there were development funds for those versions but... $79 paid one year ago per user simply won't cut it. We could factor backward and see where they'd have to cut off development.

(Personal assessment: I think if more people did subscribe we would see more support of previous "key" versions. The subscription model supports a concept I consider as 'always on'. If taken to its logical conclusion A:M subscribers could have access to all websubscriptions releases simultaneously. In this way users could easy move forward and backward through versions and use whatever is optimal. It should be emphasized: I don't see enough interest in recurring websubscriptions to support this concept. Most user states default to 'always off'.)



That will never be the only solution but subscribing or upgrading is the only way to keep moving toward perfection.
As long as new features are developed, perfection will never be achieved.
But we can define perfection and measure it in degrees.
For instance, the only way to have perfect stability is for all elements of new development to stop.
This is especially problematic for programs like A:M where users can (and will!) create their own feature sets out of what is currently 'on'.

Each of us has to make informed decisions regarding our projects.
Hash Inc has spend considerable time and money trying to find a way to keep development going while maximizing stability.
While not perfect, I think they've got it.

Subscribing each year may not be the best solution for everyone but its the best solution for 'all-ways'.

(My take)

This is what has griped Martin for years. New Features and Stability are goals at odds with each other. New Features cannot be quickly introduced in software as complex as A:M without the risk of some side effect. Martin is of the opinion that A:M needs no new features, btw.



My personal sense is the new issues are increasingly small. Only when I started doing far out stuff like cloth did I have trouble.

My other sense is multi-core users have more trouble than I do. But Steffen is on that.







Nope.


A programmer at the telecom where I used to work claimed the only completely 100% bug-free software is the one that launches the Space Shuttle. Not the software the handles your bank account, not the software that handles US air traffic, certainly not the software that handles the baggage on those flights. Not Maya, not XSI, not 3DS Max. A:M is at least as complex as baggage handling.

The Shuttle Launch program has had hundreds of millions spent developing and testing it and trying to break it and retesting it.

Consumer software like A:M depends on user feedback to find hidden bugs. (BTW, because of that, I prefer the frequent updates we have now to a once-a-year scheme where you have to wait a year to get something fixed)

I suppose that sounds unfair. You paid your money and it ought to be perfect right out of the box, right? But A:M is not a bar of soap, or a hula-hoop. Perfection can't be had for the money involved. A:M sells for less than plugins for the other programs do. If perfect is the only acceptable level then the only choice is to not sell the software at all.



I don't believe a "16" is planned right now. I think 15 will be supported and incrementally improved for the foreseeable future.

That has always been the case... after a new version was introduced the old one was retired and not bug fixed further. The current rolling subscriptions eliminate that problem.

I have 7 A:Ms installed on my PC. Don't need them, really. But I have them just in case. Occasionally I run v11 or v13 to see if something rendered differently way back when.

Did I mention perfection will never had?

You know you can have multiple versions installed, right? Si-mull-tane-ee-us-lee.

If 15f was great for you, use that for your everyday stuff until another supercedes it in reliability. And if you should happen to need something that is only in g or h or i... fire them up.

I have saved every installer I've used since 2003, just in case I need to roll back for some reason. That's 83 of them. A few months ago a guy came here with some v7 model that wouldn't load right in V15. But I was able to resave them in v10.5 and then they worked for him in V15. Glad I still had 10.5 around. Wouldn't want to use it every day though.

Robert hits a lot of good points and I'd like to expand on one:



Version numbers are something of a necessity (for programmers) but for all intents and purposes to A:M Users they are gone.
This syncs with my previous thoughts regarding A:M as a program that is always on.
Of course we can still take a slice out of time and use that via the current A:M CD. That is still an option.

As for me I'm ready to move on.
I'll soon be subscribing to A:M 2010!

Thanks for the info guys.