Rules

This page is generated from the code and must not be edited by hand.

Ingence concludes nothing without a rule. This is every rule the engine ships, in the Rulebook at revision 4 (digest 36ecb635f8fde81c2c68bc9ad844de3d): 114 rules in force. Each page gives the rule’s sentence in plain words, what it concludes, what it may be tuned by, and what would defeat it.

The rules are listed in the order the engine applies them. Under each heading a rule’s sentence is the one the Dashboard leads with.

What the walk itself records

Records a fact while the folders are walked or a container is opened. Run by the walk, not by the reading of the rules. There are 4 of them.

Rule	In plain words	Concludes	Available in this version
Cascade container member	A ZIP or other container holds only what opening it found, and the Index notes whether opening finished, stopped at a limit or was refused; an archive link is read from its text and never opened.	embeds	yes
Census directory aggregate	A folder holds exactly the files, folders and bytes the walk counted inside it and beneath it.	contains	yes
Census filesystem containment	A folder holds every file and folder the walk found directly inside it.	contains, named directory_of	yes
Census object kind	Anything on the disk that is not a file, a folder or a shortcut is recorded as what it is, such as a pipe or a device.	a recorded fact	yes

What files are

Says what a file or a folder is, by placing it in a class. There are 31 of them.

Rule	In plain words	Concludes	Available in this version
Acquisition stage from format	A file in an image or electrophysiology format that an instrument writes is placed at the recording step, unless its contents contradict its extension or its own name says it was processed.	a class	yes
Analysis stage from derived sense	A file whose name says it holds a trajectory, which is worked out from recordings rather than recorded, is placed at the analysis step, unless it is a program.	a class	yes
Condition from quantities	A file whose name carries an amount, such as 10uM or 37C, is taken to record the condition it was made under.	a class	yes
Condition inherited from folder	Not available in this version. A file whose name states no condition takes the condition its folder’s name states.	a class	no
Date neighbourhood of a file	A date in a file’s name, or else in the nearest folder name above it, says roughly when it belongs, never exactly when it was made; two dates make a span.	a class	yes
Date neighbourhood of a folder	A date in a folder’s name says roughly when what it holds belongs; two dates make a span.	a class	yes
Declared digest contradicted	A file whose bytes the engine read under the same checksum method its archive used, and found a different checksum, does not hold what the archive said.	a class	yes
Declared sha256 contradicted	A file whose bytes, read here, give another SHA-256 than its archive declared is marked so, and the reading here is the fact.	a class	yes
Earliest credible time	Each file is given the earliest time it is known to have existed: a time written inside it, where one was read, and otherwise the earlier of its creation and its last change, since a copy is created after it was last changed; an impossible time, such as one from 1969 or one later than the scan that saw the file, is set aside; a document’s creation date that is really the date of the blank template it was made from (midnight, years before the document was saved) is kept but not believed; where the time came from is kept with it, and a stand-in file is left out.	a class	yes
Empty replicate is no data point	An empty file sitting among files of the same kind that hold data, such as one movie’s empty tracking table beside the others’ full ones, is taken as no data point and is left out when an analysis is weighed.	a class	yes
Experiment folder from contents	A folder is taken as one experiment when it holds an instrument’s raw files together with files named after them, such as drift-corrected copies and results tables, and no folder above it holds the same ones, or as one session’s raw files when it holds only those, or as an analysis when it holds only files named after raw files kept elsewhere, or as an analysis when it holds one program’s many folders side by side, one per movie or cell, each with the same few files, all without reading what any word means.	a class	yes
Experiment folder from dated name	A folder whose name holds a date beside other words is taken as an experiment, unless a folder above it already has such a name, or it is one of many folders a program wrote, one per movie or cell.	a class	yes
File role unless program	A file is taken to serve the purpose its name states, such as a protocol or notes, unless it is a program or sits inside an installed program.	a class	yes
Filesystem noise from spelling	Files a computer makes on its own, such as .DS_Store or Thumbs.db, are recognized by how their names are spelled and set aside; an office program’s lock file, whose name starts ~$, is left alone, because it means somebody was at work.	a class	yes
Folder copied by declared clocks	A folder whose files an archive says it received together, on one day, was copied there whole on that day; the archive’s time is the copy’s, never when the files were made, so nothing is put in order by it.	a class	yes
Folder copied in bulk	A folder whose files were all created within minutes of one another, after they were last changed, was copied here whole on that day, which explains its creation times and asks nothing.	a class	yes
Format from extension unless contradicted	A file is taken to be in the format its extension names, when that extension means one format and nothing else, unless its contents begin the way another format does; a container the format is built on, such as HDF5 or ZIP, does not count.	a class	yes
Later stage from name	A file’s name says it is past the point of recording when it has a word such as processed, filtered, stitched or merged in it; instrument software saves such files in its own format too.	a class	yes
Miscellaneous files approved	A folder someone kept as their miscellaneous files is claimed so, and nothing in it is moved into an experiment.	a class	yes
Naming is authored or emitted	Not available in this version. A name is taken as typed by a person when it holds a word a machine would not write, and as written by a machine when three or more names beside it differ only by a counter.	a class	no
Note written as a file name	An empty text file whose name is a sentence is a note someone left, and its name is what it says, so it is never counted as an empty file.	a class	yes
Organising folder from name	A folder named for analysis, summaries, exports, loose files or in-between steps is taken to organize work rather than to hold one piece of it.	a class	yes
Pipeline stage from name	A file is placed at the step of the work its name states, such as raw, processed or analysis, unless it is a program or sits inside an installed program.	a class	yes
Placeholder from entry kind	A shortcut, or a file that claims a size but takes up no room on the disk, is taken to stand in for contents kept somewhere else.	a class	yes
Program from extension	Files whose extensions come from programming or macro languages, such as .py, .R or an ImageJ macro, are recognized as programs, unless their contents say otherwise.	a class	yes
Replicate from indexed label	A file is taken to be the repeat its name numbers, such as rep2, unless it is a program.	a class	yes
Roi archive from members	A ZIP is recognized as a Fiji set of regions of interest when a look inside found that every item is a .roi file beginning the way Fiji’s regions do, unless the ZIP is a stand-in.	a class	yes
Series member from index	A file whose name carries a numbered label, such as cell3, is taken to be one member of a series rather than a thing on its own.	a class	yes
Stage inherited from folder	Not available in this version. A file whose name states no step of the work takes the step its folder’s name states.	a class	no
Technique from name	A file is taken to be about the techniques its name mentions, when its name mentions at least two.	a class	yes
Version from name	A file whose name carries a version, such as v2, is taken to be one version of something rather than the whole of it.	a class	yes

Which files go together

Says which things count as the same in some respect, and so forms groups. There are 23 of them.

Rule	In plain words	Concludes	Available in this version
Archive link copy of a file	A link an archive left in place of a file is a copy of the file of its name held elsewhere in the folders read, matched by name and by the folders its text names; nothing is opened.	equivalent, in respect of archived_copy	yes
Copies a person answered	A set of files someone called copies is grouped as copies, and a set someone called unrelated is no longer called copies.	equivalent, in respect of copies_a_person_answered	yes
Copies by sampled digest	Files whose contents give the same fingerprint from a quick sample are grouped as holding the same bytes, as far as a sample can tell, unless a file’s contents could not be read.	equivalent, in respect of sampled_content	yes
Experiments approved together	Every file beneath the experiments of a proposed group someone approved is grouped with the rest, so one yes covers the whole group.	equivalent, in respect of approved_experiment_group	yes
Group resolved across experiments	A set of words someone approves groups every experiment holding all of them, each still its own; every set asked is one some experiments share exactly, never a date or a number.	equivalent, in respect of resolved_group_across_experiments	yes
Group resolved within experiment	Once someone approves a set of words as a group, every experiment holding that set becomes that group, each experiment (each day) its own.	equivalent, in respect of resolved_group_within_experiment	yes
Group series across experiments	Once someone approves a set of words as a group, every experiment holding it stays in that set’s series across the days, even when it also holds a more specific approved set.	equivalent, in respect of approved_group_series	yes
Identical by content digest	Files whose whole contents give the same fingerprint are grouped as holding the same bytes.	identical	yes
Identical by declared digest	Two files whose archive gives them the same checksum hold the same bytes, as the archive says it; a step below what the engine proves by reading the bytes itself, and taken back where it reads them and finds otherwise.	identical	yes
Identical by declared sha256	Two files whose archives declare one SHA-256 hold the same bytes, unless the engine’s own reading of either disagrees.	identical	yes
Name family without stages	Files whose names share every word, leaving out words for the step of the work such as raw and processed, are grouped as one family of names, unless one of them is a computer’s own file or a stand-in.	equivalent, in respect of name_family	yes
Named for a series	A run of numbered files in one folder (hela_001 to hela_010) and a table beside them named for their shared name with a word saying it holds results or a summary (hela_results.xlsx) are grouped together.	equivalent, in respect of series_summarised	yes
Named for an acquisition	A recording from an instrument and every file whose name contains the recording’s whole name, such as a table a tracking tool named after the movie it read, are grouped together, each file going with the nearest recording of that name in its experiment; so is a file whose whole name, ending included, other files’ names contain (an image viewer’s “movie.tif - C=0.tif”), even when it is no recording, and an image with the label, mask or annotation named after it in its folder (“vol07_labels.nii” beside “vol07.nii”), whatever step the image is at.	equivalent, in respect of acquisition_named	yes
One name in sibling folders	Files with exactly the same name in folders side by side under one folder (images/a.tif and masks/a.tif) are grouped together, unless the name turns up in so many side-by-side folders that it is just how each folder names its files.	equivalent, in respect of one_name_beside	yes
Replicates of one condition	Files in one folder whose names differ only by a counting number, such as _001 and _002, are taken as repeats of one condition; amounts with units, such as 10mM, must match in the same places, and a movie and its table are kept apart.	equivalent, named same_condition_as, in respect of experimental_condition	yes
Same array shape	Stored arrays with the same shape, element type and storage order, in two dimensions or more, are grouped as sharing one shape.	equivalent, named same_shape_as, in respect of shape	yes
Same column schema	Tables with the same column names in the same order and the same delimiter, two columns or more, are grouped as having the same columns, which says nothing about being one experiment.	equivalent, named same_schema_as, in respect of schema	yes
Same composition	Two folders are counted as alike when the kinds of things directly inside them match and each holds at least three.	equivalent, in respect of composition	yes
Same experimental condition	Files whose names give the same amounts and the same condition or reagent words are grouped as one experimental condition.	equivalent, named same_condition_as, in respect of experimental_condition	yes
Same name identity	Files in one folder whose names match in every word, number, label, amount, date and version, with the numbers in the same places, leaving out words for the step of the work and for repeats, are grouped as one named thing, unless one of them is a computer’s own file or a stand-in.	equivalent, in respect of name_identity	yes
Same name identity within experiment	Files of one experiment whose names match in every word, number, label, amount, date and version, leaving out words for the step of the work and for repeats, are grouped as one named thing even when they sit in different folders for different steps (raw, processed, analysis), but not across folders named for different days or repeats; outside any experiment, only within one folder.	equivalent, in respect of name_identity	yes
Same observed metadata outside programs	Files that match in name, size, last change and kind are grouped as probably copied from one another, except inside an installed program; whether they hold the same bytes is a separate question.	equivalent, in respect of observed_metadata	yes
Same version family	Folders whose names differ only by a version mark are grouped as versions of one thing.	equivalent, in respect of version_family	yes

Pairs suggested for a look

Proposes that two files are related in one of the eight ways, for a person to look at. A proposal is never a conclusion. There are 18 of them.

Rule	In plain words	Concludes	Available in this version
Concerns a shared technique	Not available in this version. Two files are suggested as being about the same technique when their names mention at least two techniques in common.	concerns, named about_technique, in respect of technique	no
Converted from sibling format	Not available in this version. A file is suggested as a conversion of another holding the same content in a different format, when their names belong to one family.	derives, named converted_from	no
Derivation by a scripts calls	A file is suggested as made from another when one script’s own lines read the second by name and write the first by name, which says which came first; never inside an installed program.	derives, named processed_from	yes
Derivation named by a describing file	As shipped, suggests nothing: set to read the time, two files a script, a notebook, a log, a protocol or notes mentions would be suggested as one made from the other, the earlier first.	derives, named processed_from	yes
Derivation named for an acquisition	A file named after a recording, or after another file’s whole name, ending included, is suggested as made from it, or from the file whose whole name its own holds, whatever step its name states, and never from another file named after it by time alone; never inside an installed program.	derives, named processed_from	yes
Derivation of one name in sibling folders	Of two files with the same name in side-by-side folders, the one in a folder whose name means raw data, images or input is suggested as the source of the one in a folder whose name means labels, masks, ground truth or output; never inside an installed program.	derives, named processed_from	yes
Derivation summarising a series	A table named for a run of numbered files beside it is suggested as made from each of them, one link a file, and no file of the run as made from another; never inside an installed program.	derives, named processed_from	yes
Derivation within name identity	Two files with one name in two folders of an experiment, but for words naming a step of the work, are suggested as one made from the other, the earlier as the source, link by link, whatever step each name states; never two files of one folder, nor inside an installed program.	derives, named processed_from	yes
Derives from name family	Not available in this version. A file is suggested as made from another when their names belong to one family, the first is raw data, an image made from data or a table, and at least two separate kinds of evidence agree.	derives, named processed_from	no
Figure renders from tabular source	Not available in this version. A figure is suggested as drawn from a table of measurements or a summary whose name belongs to the same family.	derives, named rendered_from	no
References a mentioned file	A file is suggested as pointing to another when a path it mentions leads to that file, in its own folder or elsewhere in the data, except inside an installed program.	references	yes
Related by a describing file	Two files a script, a notebook, a log, a protocol or notes mentions are suggested as related, with which, if either, was made from the other not known; never inside an installed program.	concerns	yes
Related by one name in sibling folders	Two files with the same name in side-by-side folders whose names say nothing about which came first (left/a.tif and right/a.tif) are suggested as related, with neither said to be made from the other; never inside an installed program.	concerns	yes
Related within name identity	Two files with one name in two folders of an experiment, but for words naming a step of the work, that no time puts in order, are suggested as related, with which, if either, was made from the other not known; never inside an installed program.	concerns	yes
Replicates by design	Not available in this version. Two pieces of work are suggested as repeats of one another when their names give the same amounts and words and at least three separate kinds of evidence agree.	equivalent, named replicate_of, in respect of experimental_condition	no
Succeeds by version token	Not available in this version. One folder is put after another in a series of versions when both names carry a version number and they are versions of one thing.	succeeds, named version_of	no
Summary derives from its group	Not available in this version. A summary is suggested as made from the files of its group, when the group holds three or more files under one condition.	derives, named summarized_from	no
Varies in one respect	Not available in this version. Two pieces of work are suggested as variations of one another when their names share every word, they differ only in their condition, and at least two kinds of evidence agree.	equivalent, named variant_of, in respect of design	no

What is raised for attention

Raises something for a person’s attention without concluding anything. There are 31 of them.

Rule	In plain words	Concludes	Available in this version
Access failure is concentrated	A folder is flagged when five or more of its files could not be read, far more often than elsewhere, or when a folder inside it could not be opened at all.	a finding	yes
Binding is ambiguous outside programs	A file is flagged when a word in its name could mean two or more things equally and nothing tells them apart; inside an installed program, whose names are code, it is only marked.	a finding	yes
Broken addresses in folder	A folder holding files whose addresses on the archive answered not found or moved, or files with an address that could not be read as one, is flagged once, with how many, and nothing is said about whether those files exist elsewhere.	a finding	yes
Claim class a person answered	Where someone said a rule read a whole class of files wrongly (every file it gave one value), each of those readings is shown as in doubt, with their answer as the reason.	a finding	yes
Clocks disagree	Not available in this version. A file is flagged when two equally trustworthy times for it disagree by more than a day.	a finding	no
Comprehension failure is concentrated	A folder is flagged when ten or more of its files could not be made sense of, far more often than elsewhere, or when a folder inside it could not be opened at all.	a finding	yes
Content contradicts extension	A file is flagged when its contents begin the way one format does and its extension names another, leaving out containers such as PDF, HDF5 or ZIP that other formats are built on.	a finding	yes
Content says do not analyse	A file whose own text says “do not analyze” or “do not use” is flagged, because somebody left a note about it; nothing is skipped for it.	a finding	yes
Content unreadable	A file whose contents cannot be read is flagged, saying why: it holds nothing beside siblings that hold something (an empty output), it is empty, it is empty and untouched since before a set date, it is a link the archive left in place of the file, it is a link to somewhere outside the data or to something gone, or it is not an ordinary file; inside an installed program, where empty files are ordinary, it is only marked.	a finding	yes
Copied in bulk	A folder copied in all at once (five or more files, four in five created after they were last changed, all within ten minutes) is asked about only where its times make a later step, such as an analysis, look older than the raw data it came from.	a finding	yes
Declared attribute to clarify	Each attribute an archive’s listing states on its files is asked about once, by its name, for you to say what it means; nothing is assumed.	a finding	yes
Declared copy window exceeded	A folder whose files an archive says were copied in over more than a day is flagged, so the one-day copy window can be adjusted.	a finding	yes
Domain pack would read this pool	A subject’s vocabulary pack you have not loaded would read words or file types nothing reads here, so it is offered, with how many it would read; nothing loads until you say yes.	a finding	yes
Doubt in a folder name	A folder is flagged as in doubt when its name has a word such as wrong, failed, old, test or redo in it; a weak word, such as questionable, counts only beside another.	a finding	yes
Doubt to confirm in a folder name	A folder whose name holds only weak words of doubt, such as questionable, is flagged for you to confirm.	a finding	yes
Duplicate candidate unverified	Not available in this version. Files that look like copies of one another but whose contents were never compared are flagged for checking, except where every one sits in an organizing folder.	a finding	no
Experiments may belong together	Experiments whose folders share rare words and file names are proposed as belonging together, one question per proposed group, and nothing is grouped until someone says yes.	a finding	yes
Format built on a container	A file whose contents are a container, such as HDF5, is flagged as built on that container, and a ZIP under an archive’s extension is asked about, since only the lab can say what it holds.	a finding	yes
Group incomplete against workflow	Not available in this version. A group is flagged as incomplete when it has nothing at one of the steps raw, processed or analysis, naming the step that is missing.	a finding	no
Group resolution deviation	Every question is a set of words some experiments share exactly, never a date or a number: the most shared first, then what each adds to a set you approved; asked word by word instead, each word added or missing is its own question.	a finding	yes
Indexed label may mark replicates	A label the engine has no meaning for, numbered from 1 up to a small number and found in two or more folders, is flagged for you to say whether it marks repeats.	a finding	yes
Installed software in folder	A folder is recognized as an installed program, such as a copy of Fiji or a Python environment, from the parts it holds, a name ending .app, its program-style subfolders or a majority of program files; its files are still indexed and read, and rules about experiments leave them alone.	a finding	yes
Miscellaneous files to approve	A folder of files that is no experiment and holds none is kept as it is and offered as your miscellaneous files, for you to approve.	a finding	yes
Replication below expected	Not available in this version. A group of files under one condition is flagged when it holds fewer than three.	a finding	no
Repudiated by name	A file is flagged when its name says not to use it, with a word such as ignore, failed or contaminated.	a finding	yes
Roi archive named but unlisted	A ZIP whose name says it holds regions of interest is flagged until a look inside lists what it holds.	a finding	yes
Software outside the catalogue	A folder laid out like a program’s that no program Ingence knows matches is asked about: is it a program or not?.	a finding	yes
Stage disagrees with a derivation	Files are raised together when their names or formats put them earlier in the work than the files they were suggested as made from, one question for each folder and pair of steps.	a finding	yes
Stated count may mark replicates	A name stating a small count, such as N=3, is flagged: on a program or on a file past the point of recording it is read as how many things the file rests on, and on a recording, or wherever the step is unclear, as which one of several it is, for you to confirm.	a finding	yes
Unbound word in a kinds place	A word the engine has no meaning for is asked about, once for all the places it stands, where it is written in the same part of three or more names of one folder written alike as at least two words the engine knows of one kind, which are at least three quarters of the words it knows there; that kind is only a hint, and often a wrong one, so you say what the word is.	a finding	yes
Vocabulary unread in folder outside programs	A folder is flagged when more than four in five of at least twenty words in its files’ names mean nothing the engine knows; inside an installed program, whose words are code, it is only marked.	a finding	yes

Limits every conclusion keeps to

Limits what any conclusion may say, by naming what cannot be true together. There are 7 of them.

Rule	In plain words	Limits	Available in this version
Candidate generation is bounded	At most 25 possible partners are suggested for any one file, and a file whose list was cut short is marked as such.	a limit	yes
Derivation endpoints hold roles or a scripts calls	A program is never considered as made from a file, nor a file as made from a program.	derives	yes
Derivation follows the clock	Not available in this version. No file is counted as made from one written more than a minute after it, when both carry a trustworthy time.	derives	no
Derivation is acyclic	Not available in this version. No file is ever counted as made from itself, however long the chain of files in between.	derives	no
Identity conflict blocks a derivation	Two files whose names give different subjects, samples, conditions, cultures or sessions are never counted as made one from the other, whatever else they share.	derives	yes
Identity requires a content digest	Not available in this version. Two files are said to hold the same bytes only when a fingerprint of their whole contents matches, never because their names, sizes or times agree.	identical	no
Order holds within one series	Not available in this version. Files are put in order only when they already belong to one series under the same condition.	succeeds	no

Rules not available in this version

19 of the 114 rules are declared and not available in this version. The engine does not run them, so they conclude and limit nothing today. They are marked in the tables above and each says why on its own page.

The summary the engine prints after reading the rules says that 15 rules are not available in this version. It counts the rules it evaluates, and the 4 limits in this list are applied while pairs are proposed, so it does not count them.

• Condition inherited from folder

• Naming is authored or emitted

• Stage inherited from folder

• Concerns a shared technique

• Converted from sibling format

• Derives from name family

• Figure renders from tabular source

• Replicates by design

• Succeeds by version token

• Summary derives from its group

• Varies in one respect

• Clocks disagree

• Duplicate candidate unverified

• Group incomplete against workflow

• Replication below expected

• Derivation follows the clock

• Derivation is acyclic

• Identity requires a content digest

• Order holds within one series

Rules by the relation they conclude

Of the eight relations, these are the ones a rule here concludes or proposes. The other rules conclude a class, a group, a limit or a finding.

concerns

• Concerns a shared technique

• Related by a describing file

• Related by one name in sibling folders

• Related within name identity

contains

• Census directory aggregate

• Census filesystem containment

derives

• Converted from sibling format

• Derivation by a scripts calls

• Derivation named by a describing file

• Derivation named for an acquisition

• Derivation of one name in sibling folders

• Derivation summarising a series

• Derivation within name identity

• Derives from name family

• Figure renders from tabular source

• Summary derives from its group

• Derivation endpoints hold roles or a scripts calls

• Derivation follows the clock

• Derivation is acyclic

• Identity conflict blocks a derivation

embeds

• Cascade container member

equivalent

• Archive link copy of a file

• Copies a person answered

• Copies by sampled digest

• Experiments approved together

• Group resolved across experiments

• Group resolved within experiment

• Group series across experiments

• Name family without stages

• Named for a series

• Named for an acquisition

• One name in sibling folders

• Replicates of one condition

• Same array shape

• Same column schema

• Same composition

• Same experimental condition

• Same name identity

• Same name identity within experiment

• Same observed metadata outside programs

• Same version family

• Replicates by design

• Varies in one respect

identical

• Identical by content digest

• Identical by declared digest

• Identical by declared sha256

• Identity requires a content digest

references

• References a mentioned file

succeeds

• Succeeds by version token

• Order holds within one series

Kinds of reasoning

• entailed: A conclusion cannot be wrong if what it read is right.

• defeasible: A conclusion is usually right, and is retracted when a rule it names concludes against it.

• abductive: A conclusion is the best explanation of what was seen, and another explanation could replace it.

• testified: A conclusion is stated by a person, and outranks every rule within its scope.

Rules that ship switched off

1 rule is built to run and ship switched off, because whether they hold is for a person to decide. Each page says how a project turns the rule on.

Rule	In plain words	Concludes
Succeeds by series index	One file is put after another in a series when both are repeats of one condition and the second carries the next number, but repeats whose names differ in more than one part, whose numbers tie or do not count, or whose clocks run against the numbers are not put in order, and the rule is off until a project turns it on.	succeeds

Rules replaced or switched off

16 earlier rules are no longer in force. They stay in the Rulebook’s history, so a past run can still be replayed against them.

Rule	Why it is not in force
authored_name_doubt	replaced by a newer rule
binding_is_ambiguous	replaced by a newer rule
derivation_endpoints_hold_eligible_roles	replaced by a newer rule
derives_from_declared_reference	replaced by a newer rule
extension_contradicts_content	replaced by a newer rule
file_role_from_name	replaced by a newer rule
filesystem_noise_from_name	replaced by a newer rule
format_from_extension	replaced by a newer rule
projection_derives_from_acquisition	replaced by a newer rule
raw_data_from_acquisition_format	replaced by a newer rule
references_a_declared_output	replaced by a newer rule
same_acquisition_event	replaced by a newer rule
same_name_family	replaced by a newer rule
same_observed_metadata	replaced by a newer rule
stage_from_name	replaced by a newer rule
vocabulary_unread_in_folder	replaced by a newer rule