Re: Time Estimation Question

[Donald Le Vie <dlevie -at- VLINE -dot- NET>]

I'll have to agree with Mary and others that testing documentation with
associated software is an integral part of the job, though in my opinion,
you are still performing a "usability" test of sorts...making sure the
documentation matches with software functionality is a test of how well each
correlates with the other, which in my experience is a measure of usability
(no, not the true usability testing that conjures up images of human lab
rats working with computers in a room that's wired for audio and video).

But if you're getting pushback from your boss about how you are using your
time for this project, you've got to provide your boss. Once your boss sees
how the testing portion of the project consumes such a significant portion
of your time, I think presenting some alternatives (including an estimate to
completion without the testing procedures) is in order. Up to 3 hours to
perform a single test? Doesn't that fact in and of itself say something
about the eventual usability of the product?

I can see how the range of test times (20 seconds to 3 hour) makes it
difficult for you to provide even a ballpark kind of time estimate. Mary's
suggestion of having someone other than you (the writer) perform the testing
looks like the best alternative.

Donn Le Vie
> -----Original Message-----
> From: Mary Deaton [SMTP:m_deaton -at- KWARE -dot- COM]
> Sent: Sunday, July 04, 1999 9:28 PM
> To:   TECHWR-L -at- LISTSERV -dot- OKSTATE -dot- EDU
> Subject:      Re: Time Estimation Question
>
> I think this is the most important thing to happen in testing
> documentation,
> and has nothing to do with usability.  You are verifying that your
> documentation and the software match one another.
>
> I have usually done this sort of "testing" as I write and only do it again
> at the end to find out if anything in the software has been changed since
> I
> last tested. This works well if development is closing one or more
> features
> at a time and not waiting until the end of the process to close features
> and
> integrate the software.
>
> If development is waiting until the end, or not giving you the software,
> then I think you can legitimately argue for at least 1-2 hours per
> page/topic to verify all of your procedures.
>
> We generally have someone other than the writer do this testing, since
> writers of all types are so close to their product they often do not see
> errors.
>
> Of course, your management could reject the request of this testing time
> and
> opt to ship your documentation without having this verification done.
>
> If they do, I would get it in writing that the docs have not been verified
> against the final software, and maybe even argue to include something in
> the
> docs telling the user that there is no guarantee the docs and software
> will
> match.
>
> Mary Deaton
> President, KNOWware, LLC
> (206) 682-6113
> *       Smart User Assistance and Training http://www.kware.com
> *       Microsoft MVP for HTML Help
> *       Program Associate, Winwriters 2000 Online Conference
> http://www.winwriters.com/ohc.htm
> *       Speaker, Help Technology Conference,
> http://www.winwriters.com/htc99.htm
>
>
>
> -----Original Message-----
> From: Keith Allingham [mailto:kallingham -at- PLAINTREE -dot- COM]
> Sent: Tuesday, June 29, 1999 12:48 PM
> To: TECHWR-L -at- LISTSERV -dot- OKSTATE -dot- EDU
> Subject: Re: Time Estimation Question
>
>
>      I have to disagrwith this.
>      I would *like* to agree (that would make my job way easier).
>      However, this is verification - its an integral part of the
>      documentation production process.
>      Sizing it can be iffy, I agree, but your estimate can include cya's
>      like "each procedure will take 1.5 hours to verify, assuming
>      functional s/w"
>      - if the s/w is not functional, you can't take the
>        blame for a schedule overrun
>      - if the s/w has changed, ditto.
>      Its all about CYAs and negotiation.
>      My 2 cents,
>      Keith Allingham
>      Plaintree Systems
>
>
> ______________________________ Reply Separator
> _________________________________
> Subject: Re: Time Estimation Question
> Author:  Donald Le Vie <dlevie -at- VLINE -dot- NET>  at internet
> Date:    6/29/99 1:33 PM
>
>
> I would provide an estimate on all other services BUT the one you're
> having
> trouble with. That's more of a usability issue than an editing issue.
> Different set of requirements....
>
> Donn Le Vie
> Director, Information Development
> Integrated Concepts, Inc.
>
> > -----Original Message-----
> > From: Letty Smith [SMTP:lsmith -at- CTILIMITED -dot- COM]
> > Sent: Tuesday, June 29, 1999 1:10 PM
> > To:   TECHWR-L -at- LISTSERV -dot- OKSTATE -dot- EDU
> > Subject:      Time Estimation Question
> >
> > I have searched the archives and wasn't able to come up with a good
> answer
> > to this question. Hopefully, some of you out there have faced a similar
> > situation and can give me some advice.
> >
> > I am working on editing a set of online help documents for a software
> > product.  The editing is a comprehensive one, designed to:
> >      Fix grammar errors
> >      Improve the clarity
> >      Spell check
> >      Reformat long documents into managable sections
> >      Include screen prints of processes and buttons
> >      Provide consistent naming conventions
> >      Check that each field on the software matches the documentation
> >      Run each process that the software can do, and make sure that the
> >      software actually does it the way that the documentation says it
> > does.
> >
> > The last item is the one that is stumping me.  This is very complicated
> > software, and running any one of the processes takes somewhere between 1
> > minute and 3 hours (depending on how much previous data I have to load
> and
> > create for testing).  I have had tests take a whole day to finish, when
> > the
> > system was feeling buggy.  They have also taken 20 seconds.
> >
> > Does anyone have a good rule of thumb as to how to estimate how long
> this
> > sort of editing should take?  I'd like to be able to present my manager
> > with some guesstimation, as I'm starting to get the "gee, what are you
> > DOING with your time" looks.   Currently, there is no method being used
> to
> > estimate how long it should take, so any help would be an improvement.
> >
> > How about for this sort of edit *without* the software testing?
> >
> > Letty Smith
> >
> >
> >
> From ??? -at- ??? Sun Jan 00 00:00:00 0000=
> > =
> >
>
>
> From ??? -at- ??? Sun Jan 00 00:00:00 0000=
> =
>
>
> From ??? -at- ??? Sun Jan 00 00:00:00 0000=
> =
>
> From ??? -at- ??? Sun Jan 00 00:00:00 0000=
> =
>
>


 From ??? -at- ??? Sun Jan 00 00:00:00 0000=