The initial analysis comprises following steps and decisions:

Identifying GN3 site-specific custom procedures

There might be site-specific procedures (workflows or workarounds), based on custom solutions (custom SQL triggers or stored procedures, or direct database access) that are not available in GN4. Every such item has to be analyzed to find a solution suitable for GN4.

Identifying any GN3 data not worth to be migrated

In any production database there will be number of unused items (old articles or edition templates nobody knows why they're not purged; unused desks and baskets; unused fonts, layouts, formats etc). Identify them, create a fresh backup of a GN3 database, remove the unused items and then export.

Identifying and describing data that you will re-create manually

These are automatic copy flow paths, wire baskets and Picture desk configuration and procedures, directory styles, sort orders, search dialogs, any Quick Prompts custom dialogs, extended queries, text estimating, personal preferences, personal workspaces (GN3 term: desktops), scripts, Wire, Picker, Topic queues and configuration, custom Perl/VB/SQL scripts.

Prepare the detailed description on how they're designed and port them manually one-by-one in the GN4 configuration.

Identifying and describing data that you will remap manually

These are the library objects with auto-tool, contains ids of GN3 formats and GN3 types - the correct formats and types mapping will not be migrated, so you will have to adjust all exported library objects with format auto-tool to the GN4-related format and type IDs before importing them in GN4. See About exporting auto-tool in object libraries.

The same applies to the inline pictures in articles, pointing to pictures with GN3 IDs - when you reimport the articles and the pictures, the pictures IDs will be different. Before importing such articles, adjust the picture IDs.

Comparing schemas (differences between GN3 and GN4)

In an ideal case, the GN4 schema would be exactly the same as GN3 schema, so no adjusting would be needed at all. In a real life, there will be number of differences, for example:

•GN3 elements type or attribute names that do not exist in GN4: in that case, you may want to add them to GN4 schema before importing GN3 data to GN4, or you may want to remove them (if not needed) from GN3 exported files before importing them to GN4.

•GN3 element types or attributes that have the same purpose as in GN4, but with different names: e.g. the GN3 element streamer corresponds to the GN4 element pullquote. In that case, you may want to rename GN4 element/attribute in the schema, or rename GN3 element/attribute in exported file.

•GN3 element types that match GN4 element type names, but in different order: e.g. the GN3 element kicker may appear before element headline, while in GN4 it appears after. In that case, you may need to re-order GN3 elements, or adjust exported XML files to match GN4 elements order.

Have in mind that element type differences affect also typographical formats (if these are assigned to types), object libraries and scripts (where types may be called).

Check the layout assignment to all GN3 fonts

Sometimes the fonts that require the Dingbats or Symbol layout have the Unicode layout assigned instead. This will cause problems in GN4. Make sure that all non-letter fonts have the  Dingbats or Symbol layout assigned before exporting data.

Note

GN3Export works only for the standard GN3 database structure - it won't export from custom tables, or additional column names in standard database tables.