An interesting Mif2Go project - will this work?

[Jeremy H. Griffith]

On Fri, 2 Aug 2013 08:57:56 -0400, John Sgammato <john.sgammato at actifio.com> wrote:

>I am using FM11 unstructured with Mif2Go to populate a SalesForce
>Knowledge knowledgebase. SalesForce Knowledge is a relatively new
>offering from SaleForce, much better than their older knowledgebase
>product. Ultimately this will involve hundreds of individual topics so
>I want to automate as much of it as I can.
>I have talked to Rick Quatro about this, but there are some other bits
>where I'd appreciate some ideas.
>
>To populate the knowledge base:
>1. I must generate individual HTML files from the FM content and store
>them in a single folder.

That's the default result.

>2. The SalesForce importer reads a .csv file to get location and other
>info about the content.
>
>The .CSV file looks like this:
>isMasterLanguage,Title.File,datacategorygroup.Products,Channels,Language
>1,Forcing Expiry of a Backup Image,/articles/en_US/doc/ab1101973.htm
>PAS_51,sites,en_US
>1,Restoring Selected Files,/articles/en_US/doc/ab1101043.htm,PAS_51,sites,en_US

Are the first, fourth, fifth, and sixth columns the same
for all files, or do they vary too?  If they vary, is there
a rule that describes what they need to be, for your project?

>For this release, I need only the procedures; conceptual information
>can come in the future. (I have it, but I do not want to use it in the
>knowledge base until I do some more development work on it.)
>
>After I generate all the HTML files, I somehow need to get a list of
>the filenames for the third column of the .csv, and I'd love a way to
>get the Title of the topic into the second column.

We can probably write a CSV file with that info for each 
Frame chapter, then just concatenate them for the full book.

>My plan is to go through the 6 books that will get this treatment and
>apply a new marker to all of the topics that I want for the knowledge
>base.

Please don't.  Not only is it a large effort, it is not
necessary from your description above.  The guaranteed
unique filenames we create will work.  Yours would be
absolutely guaranteed **not** to be unique,  There is
a reason we warn against this.

>I know how to set the HTML file title = that marker, but I am afraid
>that in the end I will have many surplus files, and they will all have
>autogenerated filenames like ab1101973.htm.

Even worse.

>Is there a way to automatically set the output *filenames* to match
>the titles? I understand it's a Bad Idea in case I have multiple files
>with the same title, but I will be doing enough post-processing to
>catch and fix that.

There actually is; we undocumented it for a while
because of the large number of disasters reported 
to us.  There is *no way* to tell which files were 
clobbered by others until *your* users complain 
about the links going to the wrong places...

>Is there a way to  get only the topics I want and ignore the other
>content? 

How would you identify it?  (Hint: **not** markers.)

>I can start a new topic at every H1 and H2, and I can title
>the ones I want from the markers, but in the end I will have a bunch
>of files with near-identical filenames, right? Maybe I can send the
>topics with the markers into a different folder?

No, that won't work.  The right answer depends on 
your answer to the question above.

>I have some time to work on this, but not a lot - a major release will
>be upon me next week.

Plenty of time, LOL!

>I have a henchman in India who can help with the FM work, but he
>doesn't know Mif2Go yet.

This may not be the best time for him to learn.  ;-)

>I appreciate any thoughts to set up a good, solid, reproducible
>process for this.

A few answers, please, and we can probably provide one.  <g>

-- Jeremy H. Griffith, at Omni Systems Inc.
  <jeremy at omsys.com>    http://mif2go.com/