Choosing an Execution Mode¶
Run the whole Project end to end, or one Stage at a time. The choice affects how much you can influence along the way.
The setting is Migration Mode, and it appears on the Project Home Page below the Export file selection, so you choose it before the file is uploaded and extracted. Automatic is preselected and marked Recommended; the alternative is Step-by-Step. Project Settings covers this and the other Project settings.
Both modes run the same Stages in the same order, and both leave Platform Deploy to you. What differs is whether SLIM waits for you to review the deliverables or alter the focused scope between one Stage and the next.
Automatic¶
Click Extract Assets and SLIM runs one step and then the next without interruption. Extraction finishes and Export Analysis starts; when that completes, SLIM moves on to Integration Analysis for all integrations in the export file, and then Pipeline Generation, again, working through every integration it found in the export. Then it stops. Platform Deploy is always a deliberate act - see Stage 5 - Platform Deploy.
While a run is in progress, the run controls on every Stage Page are replaced by an Automatic Mode Running… badge and a Stop button. You are not locked out of the Project: you can click any of the Stages’ Stop buttons and terminate the Automatic Mode, which stops the current Stage, and prevents any future Stages from beginning automatically - see Stopping and Restarting.
While in Automatic Mode, the Activity Log streams what is happening, the documents produced so far can be read, and the Job Status page fills in as integrations complete each Stage.
Automatic suits two situations:
The integrations are straightforward and you expect the standard analysis to serve.
You have worked with this type of Export file before and already know what instructions it needs, in which case you can put them in place before you start, rather than discovering them along the way.
Two things automatic does not do, both worth knowing before you choose it:
It does not pause for you to read or adjust a result. Each Stage takes the previous Stage’s output exactly as it stands as input into the next Stage.
It does not pause for you to narrow the list. Every integration discovered in the export is analyzed and associated new assets generated. On a large export that is a great deal of time and work you may not actually want - see Narrowing the Scope.
Step-by-step¶
Nothing runs until you press a button. Each Stage Page carries its own initiation button: Run Export Analyzer, Run Integration Analyzer, Generate Pipelines. A Stage that finishes stays finished, and the Project waits.
That gap between Stages is the whole point of the mode. In it you can:
review the document the Stage produced with other stakeholders, and edit it in place where it is wrong or if data is missing from the Export file,
add instructions to the panel on the next Stage Page, so that Stage works from your correction instead of repeating the misunderstanding,
choose which integrations go forward, from the list on the Integration Analysis and Pipeline Generation pages.
Correcting the process at the end of each Stage costs far less than regenerating later. Why Guidance Matters makes the argument in full; it is the reason the Step-by-step Mode exists.
If you attached design documents¶
One difference between the modes is in what you get rather than when.
Where a Project holds design documents, SLIM scores the generated pipelines against them and writes a conformance report, CONFORMANCE_REPORT.md, alongside the pipelines.
NOTE: Owing to an existing bug, that only happens on an Automatic Mode run. In Step-by-step operation, Pipeline Generation does not produce it and no button triggers it.
If your review depends on a signed-off design, run SLIM in Automatic Mode until the defect is addressed. This is a known gap rather than a decision, and it is expected to close. Until it does, check whether the report is actually there before relying on it, and do not read its absence as a finding about the pipelines. See Supporting Documents.
Switching between modes¶
Switching between the two modes is not symmetrical, and the asymmetry is worth understanding before you choose.
Automatic to Step-by-step: any time. Press Stop on any Stage Page. The confirmation tells you what it does - Stop the automatic pipeline? You can re-run individual steps afterwards. - and it is accurate: stopping ends the run and leaves the Project working a Stage at a time. Any stop has this effect, including one during extraction. Work that had already finished is kept until or unless Stages are re-run. See Stopping and Restarting.
Step-by-step to Automatic: not from the app. Nothing on the Project Home Page or on any Stage Page turns Automatic Mode back on. The only route is Replace Export, which does return the mode to Automatic - and deletes every analysis document and generated asset in the process. That is not switching modes; it is starting the Project over. See Uploading an Export.
Treat the choice as one-way, given what switching back to Automatic Mode costs. A Project started in Step-by-step mode stays that way for the life of that Project. A Project started in Automatic Mode can become step-by-step the moment you want it to, and cannot go back.
Understanding which mode you are in¶
While a Stage is in progress, the Automatic Mode Running… badge answers the question at the top of each Stage Page. While nothing is running, nothing on screen names the mode, so you have to read it from behavior.
The clearest moment is extraction. In Automatic, SLIM takes you straight into Export Analysis and begins the process. In Step-by-Step, an Extraction complete panel appears with a Go to Export Analysis button, and nothing happens until you press it. After that, the tell is simply whether the next Stage starts by itself.
The run logs tell you as well, where Enable diagnostic logging was switched on for the Project. With that setting on, a logs folder appears in the Project Tree holding one folder per run, and downloading a folder as a ZIP archive includes them. An Automatic Mode run gathers the whole run into a single folder, named for the time it started and ending in auto, with each Stage’s own log folder nested inside it. Step by step writes each Stage’s log folder on its own, with nothing wrapping them. So the shape of what is there answers the question even after the fact. With the setting off the logs are still written, but nothing shows them to you.
One shortcut is reliable: a Project that has been stopped at least once is working step-by-step, because stopping is what puts it there.
Which to choose¶
When you start working with SLIM, use the Step-by-step Mode with a small Project you are familiar with and willing to rework later on.
Running a simple export a Stage at a time allows you to see the shape of the process, what each Stage actually produces, and where your additional guidance makes a difference to SLIM and its deliverables, before attempting to process a whole export hands-free with a single click. Once you know a given Export file’s needs, and your operating circumstances and constraints, running in Automatic Mode can save real time.
There is one caveat to weigh against that, and it follows from the switching rule above: Automatic Mode is the more flexible starting point, because you can always stop it before continuing the process a Stage at a time, whereas Step-by-step Mode commits you for the life of the export. If you want to keep both options open on a Project that matters, start in Automatic Mode and stop it at the first Stage you want to inspect.
For learning, the friction of the Step-by-step Mode is the point, so accept the commitment and use a Project you are willing to rework repeatedly.
Cost difference¶
Broadly, the work through either mode is substantially the same, and so costs the same in either mode.
Step-by-step Mode costs a little more, as each Stage is started and stopped separately where automatic pushes straight through. The difference is small.
The difference in cost is rarely the deciding factor in the choice of Migration Mode, and the reason is worth stating plainly: correcting an analysis document before generation runs costs far less in money and time than generating from a flawed analysis and regenerating afterwards.
With no refinement at all you may go through the Pipeline Generation Stage more than once before you have something useful, and Pipeline Generation is the longest and most expensive Stage. Set against that, the overhead of stopping between Stages is minimal.
What genuinely moves the number is the model and the effort level, and the options for those are the same in both modes - see Project Settings.