[Framers] FrameMaker to Confluence

[Robert Lauriston]

More signs that Atlassian is a screwed-up company: they resolve bugs
as Fixed when they have not in fact fixed them.

https://jira.atlassian.com/browse/CONFSERVER-39129

Or they resolve them as Won't Fix on the theory that if they haven't
fixed them in a few years, no one cares.

https://jira.atlassian.com/browse/CONFSERVER-39146
https://jira.atlassian.com/browse/CONFSERVER-39140

On Wed, Apr 18, 2018 at 12:14 PM, Robert Lauriston <robert at lauriston.com> wrote:
> 1. Word is the only supported import format.
> 2. If you use the standard Heading1 etc. it can break things up into
> topics, but subtopics will be sorted alphabetically.
> 3. Auto-numbered paragraphs are converted to literal text.
> 4. Cross-references are lost.
>
> It's bizarre to me that Atlassian doesn't recognize how much business
> it must lose by making it so hard to migrate, but there are lots of
> other signs that the company's not run too well.
>
> I saw some posts somewhere from someone who built an in-house
> integration tool using the Confluence SDK or REST API or something
> like that.
>
> On Wed, Apr 18, 2018 at 11:50 AM, Shaw, Brian
> <Brian.Shaw at dieboldnixdorf.com> wrote:
>> Hi Framers,
>> I'm hoping somebody has had some experience in moving FrameMaker books into Confluence. Out-of-the-box, Confluence only supports import from Word docs, so the approaches are to either save as RTF and then to Word, or save as PDF and then to Word. PDF appears to be the better intermediary format as we can import a complete book at a time, but maintaining the formatting during the import is approximate at best.
>> Numbered lists are particularly problematic, and strange things happen to graphics and tables. The best option we've been able to find so far is to open PDFs directly in Word and save as docx. This preserves the numbered lists after the Confluence import, except that the extras embedded in a list, such as graphics, notes, etc. break the numbering. A soft return can fix things by joining the extras to the numbered paragraph above, but the process is clumsy. Other issues are that we lose all of our cross references and special formatting such as monospaced fonts for code examples. There doesn't seem to be any options for mapping the formats during the import.
>> Another issue has been PDFs that were used as graphics in the source FM files. While they look great in an FM to PDF save as, Confluence doesn't know what to do with them. We have a workaround by initially saving the PDF images as PNGs. We then save each of the FM docs as MIF, replace .pdf with .png in the MIF text, and then resave the MIF as FM, but that's also long-winded. If somebody knows of a way to get Confluence work with PDF images, that would be cool.
>> As background, we're using Frame 2015 for general production, but some of the team also have 2017. We also have the latest versions of Acrobat and the Office products.
>> Thanks in advance,
>>