Update on IEC 62304 revision - May 2012 - 2013 (TBD)

[Marcelo Antunes]

Hello all

ISO TC 210 JWG 2 met this week to further discuss the revision of IEC 62304 (and the creation of IEC TR 80002-2).

We've finally managaed to come to mature the working draft enough to circulate a first committee draft.

The main idea behind the revision is to clarify some confusing parts of the standard, for example, to clarify software safety classification issue.

Another interesting development was the addition of an annex to deal with legacy software (software created before IEC 62304 was created and still used today). It follows the same idea of the legacy annex to the usability standard (62304) to permit that manufacturrs follow an alternative process when applying IEC 62304 to legacy software (otherwise all software would need to be redesigned to meet the standard :-)).

For those which can have access thru the National Committes, expect a CD in the following months.

[sagai]

One of the improvement would be good to see once as regard to this standard is to remove the sequential determination of tasks and to really enable its usage in a non waterfall environment.
I would really challenge if anyone is indeed conforming with this standard on an implementation level in the industry, anyway.

[glork98]

In Reply to Parent Post by sagai

I would really challenge if anyone is indeed conforming with this standard on an implementation level in the industry, anyway.

Doing my best, every day!

I think 62304 can be followed with an agile process but re-doing the requirements, architecture and design docs frequently would be a big hit on productivity. It does work best for projects that follow a waterfall approach.

For clean-sheet products where waterfall is the best approach (IMHO) following 62304 is quite natural.

[sagai]

It is hard to believe if there is any project recently can afford to do a waterfall life-cycle.
Do you know any?
I am not referring to the ones pretend to be compliant by management and quality folks, but when really, I mean really no coding occurs prior having all the requirements defined in accordance with the flavor of 62304.
I have seen enough fake implementation saying, oh yes, it is 62304 compliant, however and of course, it was very not the one on the developers level.

I seriously do not believe anyone can survive without continuously deliver running software to customer during the design phase in order to validate it and this is actually not really the way defined in a sequential, hierarchical model of 62304 (I know it is defined any model is okay, but for me it just does not really support other than the waterfall life-cycle).

But I am open to learn and see examples, that is one of the reason I spending my time here sometimes.

[Gert Sorensen]

I am very much looking forward to the draft being available to industry in general, especially because of the clarification of software classification. Too bad though, that it will require access through the national committees. Will it be made available for comment by industry at a later stage?

[Marcelo Antunes]

One of the improvement would be good to see once as regard to this standard is to remove the sequential determination of tasks and to really enable its usage in a non waterfall environment.
I would really challenge if anyone is indeed conforming with this standard on an implementation level in the industry, anyway.

There´s no suggestion to change anything related to this matter at the moment however...

IEC 62304 does not require an waterfall lifecycle model, nor a sequential determination of tasks. Not sure why you said that those are requirements.

[Marcelo Antunes]

I am very much looking forward to the draft being available to industry in general, especially because of the clarification of software classification. Too bad though, that it will require access through the national committees. Will it be made available for comment by industry at a later stage?

No, standards development are always performed by WG members and participating NCs. Sometimes (case of the revision of ISO 13485) some input from outside the committees is wanted, but it´s only informal in the way that it´s not included in ISO/IEC rules.

Just as a curiosity, why do you think it´s bad? The idea is that the industry needs to participate in standards development, but needs to do it thru the NCs (which in fact are the only members of ISO/IEC).

[Gert Sorensen]

In Reply to Parent Post by Marcelo Antunes

Just as a curiosity, why do you think it´s bad? The idea is that the industry needs to participate in standards development, but needs to do it thru the NCs (which in fact are the only members of ISO/IEC).

Unfortunately, not all of us in industry has a seat with the NC
So, basically, we don't have any idea what the NC is working on (at best we have an idea of the overall scope, but definitely not a more detailed knowledge of what the focus of the WC is). If you - as an industry professional - have a natural interest in what changes are coming within your field of expertise, then it is a little bit sad that you can't have access to some of that information. Just my