Re: Zipper , an example (software testing)

[andy butler <akbutler@tiscali.co.uk>]

Charles Zwicky wrote:
> 
> 
> Andy,
>  I think this assesment is too cynical. 

sure, :-) fit's the facts though.

Where are the bug free DSP devices? 
( other than simple ones )

> I have done beta testing with 
> several pro audio manufaturers (I am a professional recording engineer) 
> and the reality is that the "programmer " is often a part of a team of 
> developers and programmers, and in DSP driven products, there are 
> programmers dedicated to user interface, DSP coding, algorhythm design 
> API integration (if it's software) etc.

Right, and that would be the case with Roland.

Actually I don't think Roland have a lot of bugs in their gear.

I'd be fairly certain they have guys to systematically test their products.
( but who wouldn't conceive what LDers would want to do with it)


>  As a beta tester, I was perhaps 
> the only member of the 'team' who dealt extensively and exclusivly with 
> the end product.
> The basic reality is that the alpha and beta devices 
> are broken and malfunctioning and need sombody with excellent 
> troubleshooting -and particularly communication - skills to sus them 
> out.

wanna job ?   :-)

>  At some point, there is a 'drop dead' date and the thing gets 
> released, warts and all.

Well, I'd agree with that.

That's why the LP-1 thing works so well.
There's no cut off date.
The buyer is a tester, and possibly a contributor.
...and generally buyers end up happy.

The same model works for software loopers.

andy butler

> 
> -CZ
> 
>> Charles Zwicky wrote:
>>>
>>> It really falls on the Beta testers. If they don't use the devices 
>>> the way that us on the cutting edge might then these things slip past 
>>> them.
>>
>> Roland knew about some of the RC-50 issues already.
>> Don't think they have much of a clue really, apart from being top 
>> notch at profits.
>>
>> Complex loop devices really need something more than to rely on the 
>> poor old "Beta Tester", as in someone who plays their music in the way 
>> they like to see if it works out on the new device.
>>
>> Indeed, it's highly desirable to have a beta tester like Andre Lafosse 
>> who'll push the technology to it's limit, but  how realistic is that 
>> hope?
>>
>> As cpr hints, it's either down to the programmer to ensure that 
>> there's no way to crash the device for any possible sequence of button 
>> presses, or they should get someone else to do it. Typically, the 
>> programmer isn't the best tester of their own product, as they already 
>> "know to use it".
>> As looping devices increase in capability, there's 2 ways I can see 
>> this going.
>>
>> 1) Actually pay someone to test, it's a skilled job.
>> 2) Make it known that purchasers/users are "testers" right from the 
>> beginning, and have good provision for dealing with that.
>>
>>> This is why the RC-50 gets complaints here but the LP-1 and 2880 
>>> don't...and vice versa in the mainstream press.
>>
>> The reason the LP-1 doesn't get complaints here is that these are 
>> dealt with on the LP-1 forum.
>> There's not that many LP-1's in the world, so users get their answers 
>> direct from Bob Amstadt.
>> Probably the 2880 is just too simple to have a lot of problems.
>>
>>
>> Isn't the RC-50 the first looper by a major company to have these sort 
>> of problems?
>> It's a lot easier to create bug free software if all the features are 
>> known before programming starts, and I'm guessing that's how the 
>> majors do it. The independent loop developers are usually open to 
>> change their product on the fly, which is no bad thing, but eventually 
>> tends to create difficulties and complications in the software.
>>
>> If we want to be on the "cutting edge" as loopers, I think we're going 
>> to be testers :-)
>>
>> andy butler
>>
> 
>