To ‘b’ or not to ‘b’ ?

Have a problem? A question? This is the place for answers from other Express users.
Post Reply
irev
Posts: 22
Joined: 2003-03-06 15:16:43

To ‘b’ or not to ‘b’ ?

Post by irev »

I have a suggestion for future Nisus updates. What if Nisus simply added the letter ‘b’ after the new release number, for example 2.1.3b? This would alert everyone that it’s a Beta release and they should try it at their own risk. Nisus would be acknowledging upfront that it’s not perfect but they are working hard to get it that way. When it’s stable enough, it would be released without the ‘b’.

Users that are cautious or have little patience would avoid ‘b’ releases, while others willing to experiment would offer feedback to make Nisus an even better product. High levels of communication would continue between user and developer with disappointment and frustration levels lowered. “For warned is for armed.”

Other small software developers seed updates this way and avoid a lot of the negative feedback from their users. Does anyone else think this method would reduce the frustration level for both parties?

MacSailor
Posts: 290
Joined: 2003-04-03 08:38:41
Location: Linköping, Sweden

Post by MacSailor »

I think »irev» got a point here.

This could be a way to avoid bad feedback and still be able to release new versions of NWE.
Peter Edwardsson
..............................

gemboy27
Posts: 353
Joined: 2003-09-30 21:33:58
Contact:

also

Post by gemboy27 »

I also agree -- and from what I see -- when you release a new version many of the problems are related to a diverse group starts using it. That is, I have an iMac G4 with OS 10.3.5, someone else is using a earlier or older. But not just that. I use CopyPaste - X, others are


well basically getting your loyal fans to give you insight into various situations could be helpful, if people were to help

good idea Shakespeare...I mean Irev
I have never let my schooling interfere with my education/Mark Twain (1835-1910)

dshan
Posts: 334
Joined: 2003-11-21 19:25:28
Location: Sydney, Australia

To b or not to b

Post by dshan »

The problem here is what is a beta release and what is not? When does a piece of software go from being alpha, beta, etc. to live/generally available/production? How many (known) bugs, and of what severity level, should prevent a new release from being, um, released? If the answer is zero then no software will ever be released.

Nisus currently seem to do so called "internal betas" where new versions are tested by their internal people and (at least sometimes) by a select group of customers they deal with privately. These betas are not released for download by the general public and are (I presume) only "released" when the number and severity of known bugs is below some critical threshold(s) deemed acceptable for general release. Sometimes commercial software companies will do an additional round of "public" betas just to make sure they haven't got any nasties hidden in the code that were not found during the internal betas for whatever reason. This can be expensive, complex to coordinate, and delay release of the software for quite some time, always an issue when the company must make money to survive. Inevitably after release the users start finding further bugs, some serious, some trivial, some that were already known about when the new version was released (hopefully documented in release notes or READMEs when the s/w is released) and subsequent "point releases" or "patches" are then released to fix these asap. Thus 2.1 spawns 2.1.1 and 2.1.2, etc. until 2.2 or 2.5 or 3.0 is released somewhere down the track. Many, perhaps most, commercial software companies work like this, I've worked for several myself.

Other organizations, particularly open source and shareware groups, like to release beta (and sometimes even alpha) versions of new releases to the general public to get their feedback before declaring the new release GA/final, etc. Thus you see (for example) Mozilla 0.8, 1.0rc1 or 1.8a5, 1.8b and then finally, often months later, 1.0 or 1.8 is finally declared production and "released". Some shareware apps never seem to have a release that isn't suffixed by an "a" or a "b", etc. They're eternally in some form of pre-release limbo. This approach makes sense when you're a non-commercial or small-commercial developer who can't afford to employ a team of s/w testers and/or run a co-ordinated beta test program with your major users that will provide a stream of useful feedback to your engineers on bugs encountered--you use the great unwashed masses as your beta testers and avoid complaints by labelling the software as beta, release candidate x, or whatever (and plastering your website and README files with warnings). Even so when the software is finally released it still always has bugs and they still need to release subsequent fixes such as 1.8.1, 1.7.3, etc. when sufficiently serious problems are uncovered.

The question is then whether Nisus are properly testing their releases internally before declaring them available for everyone to download, or whether they should do another "public" beta for several weeks or months prior to officially releasing each version. Is a public beta going to repay the extra cost and delays in releasing each new version with a significantly reduced bug count and greater customer satisfaction at the end of the day? Will they get sufficient useful feedback from users to justify the effort and cost? (Smart users who don't like being used as guinea pigs often avoid s/w with "a" or "b" after the version number.) Are they better off just putting out regular fixes as new problems are uncovered after each new release?

Both approaches described above require fixes to be issued after the initial release, neither will ever result in perfect bug-free releases and public betas significantly delay the day when you can start making money from a new release (charging for betas always upsets people!), which has to be considered when you are in it for the dough rather than "just the love baby". It's not any easy choice.

midwinter
Posts: 333
Joined: 2004-09-09 18:07:11
Location: Utah
Contact:

Post by midwinter »

The folks at Omni have been able to use the public beta system as a way of drumming up all kinds of interest in their product.

Granted, betas in web browsers are far less likely to delete your novel than word processors, but it still might be profitable for Nisus to adopt this model.

Todd
Posts: 42
Joined: 2004-02-18 12:24:45
Location: Roaming the Universe
Contact:

Post by Todd »

Many literary scholars believe that Dostoyevski’s epic novel, “Crime and Punishment” would have reached 3,500 pages – shattering the record length for novels at the time – if only he hadn’t written it with ClarisWorks 1.1.3beta, which was prone to crashing during the long Russian winters of 1864.

charles
Posts: 481
Joined: 2002-11-26 11:40:35
Location: San Diego, CA
Contact:

Post by charles »

Hi:

Thanks for this suggestion about public betas. In fact, we tried using a public beta before with Nisus Writer Express 1.0 and we found that the process was not very helpful. While we did appreciate the added number of testers working on Express, we also found that most of the feedback we received from users was less useful such that the few really good ones tended to get lost in the shuffle. In the end, the public beta process just seemed to lengthen the time it took for us to get out a product without substantially improving the quality.

That said, we have been talking quite a bit internally here about how we can increase the quality of our releases. We plan to make some changes to how we conduct our beta testing program and other internal development procedures to improve our quality. Those changes may include some form of public or semi-public beta in the future.

-Charles
Charles Jolley
Nisus Software, Inc.

midwinter
Posts: 333
Joined: 2004-09-09 18:07:11
Location: Utah
Contact:

Post by midwinter »

Charles: I'm not suggesting that a public beta would be useful for bug fixing. I'm saying that if you ran a Nisus public beta thread over at Macnn and also at AppleInsider, you'd get all kinds of interest, I imagine. It seems like every week there's a new thread by someone who's unhappy with Word and thinks Mellel is fugly.

charles
Posts: 481
Joined: 2002-11-26 11:40:35
Location: San Diego, CA
Contact:

Post by charles »

midwinter:

What a very interesting idea! You just might find one of these appearing very soon.

Cheers,
-Charles
Charles Jolley
Nisus Software, Inc.

midwinter
Posts: 333
Joined: 2004-09-09 18:07:11
Location: Utah
Contact:

Post by midwinter »

charles wrote:midwinter:

What a very interesting idea! You just might find one of these appearing very soon.

Cheers,
-Charles
Charles,

You might consider looking at this MacNN forums thread about the Omniweb 5 public beta. Specifically, notice how active the Omni folks are in it, taking the time to respond directly to posters' concerns. You guys already have that going on over here, but I seriously think that if you started a thread there and remained active in it, it could be a remarkably effective PR move for you guys, since public betas mean interest and usually enthusiasm.

Post Reply