Document comparison and testing
ByteCompareByteCompareProve the output is right, page by page.
ByteCompare puts two versions of your documents side by side and shows every difference between them: wording, layout, fonts, colours, lines and images. AFP, PDF, HTML or text, in any combination, from a single letter to a whole print run.
Sound familiar?
What it does
AFP against PDF, PDF against PDF, two HTML emails or two text reports. ByteCompare reads each file into the same model of pages, text, lines and images, so an AFP from the old system can be checked against a PDF from the new one.
The two pages sit side by side with every difference outlined, and a list in the middle to step through. Changed wording shows exactly what was there and what replaced it.
Letters are paired by an account or policy number, from an AFP index tag or the wording on the page, and pages are lined up by their content. A letter that runs to one page fewer, or a customer missing from the run, is reported once and the rest stay in step.
Allow a little movement in millimetres, compare the wording only, or leave out colours and fonts. Ignore areas such as dates and barcodes, and wording that changes every run, such as run dates and reference numbers.
Flag each difference for change, accept it as expected, or ignore it, and add a note. When the same change appears in every letter, accept it everywhere at once.
Save the settings and use them again on the next run. A plain text report lists every difference with its status and notes, written wherever you choose, for sign-off and audit.
The same comparison runs unattended on Windows or Linux from a settings file: ByteCompare settings.bcmp fileA fileB. It writes the report and returns an exit code, so a scheduler or build pipeline knows straight away whether anything changed.
It grew out of our own migration work, where the hardest part is proving that the new platform's output matches the old. Use it for platform moves, print house changes and regression testing after every release.
A closer look
These screens use Fenmoor Mutual, a made-up insurer: six customers' renewal packs as AFP from an old system and as PDF from a new one, with a premium change, new wording, new heading colours and one customer missing, plus the same packs reprinted in smaller text.
Smaller text made one letter a page shorter. The missing page is reported once, and every later page is still compared with its real partner.
Accept the expected changes, flag the ones that need fixing, and add notes. The same change across a run is dealt with in one click.
Pair the letters by policy number, ignore run dates, and save the settings to use again, in the app or on the command line.
The same comparison in a scheduled job or build pipeline, with a report and an exit code that says whether anything changed.
How it works
Before and after, old platform and new: AFP, PDF, HTML or text, in any combination.
Pair letters by account number, set the tolerance, and ignore run dates, barcodes and anything else that should change.
Step through them side by side, accepting what was expected and flagging what needs fixing.
Save the report for sign-off, and run the same settings again on every release, by hand or on the command line.
We'll compare two of your own files with you, ideally an old and a new version of the same run, and show you what ByteCompare finds.
Built from the ground up, from real experience. ByteCompare comes from twenty years of building and migrating documents on client sites. Read how it started, or see the consultancy work we still do every day.