Single Issue Review
The single issue review is the review of one question. BABOK counts it among its seven review types and defines it in one sentence: a formal technique focused either on one issue or on a standard, in which reviewers perform a careful examination of the work product before a joint review session held to resolve the matter in focus (§10.37.3.2). Of the seven, it is the only one whose subject is a matter rather than a work product: the other six ask how closely this document has to be read, this one asks what the thing is that has to be settled. It has the ceremony of the formal reviews, planned, prepared, held in session, and the scope of a single question. What it produces is a decision on that question, and BABOK assigns nobody to write it down: that is the first thing the person who convenes it settles.
Goal
A single issue review is convened when something has to be settled and the answer has to hold. The matter is named before the session, the reviewers examine the work product each on their own and the session is held for that matter alone (§10.37.3.2). Its whole economy follows from that: the preparation is real, the session is formal and both bear on a single question.
BABOK lists seven possible objectives for a review (§10.37.3.1), and the single issue review serves two of them, both named in its own definition. Ensure conformance to a specification or a standard: the definition carries it in so many words, a formal technique focused "either on one issue or on a standard", and the singular is what makes the technique: one standard, rather than the general compliance of a product. Resolve an issue: the definition says a joint review session held "to resolve the matter in focus".
BABOK's fifth objective bundles three: answer a question, resolve an issue, explore alternatives. The single issue review takes the one its definition names. The other two are settled elsewhere, and what separates them is formality and scope rather than the objective. A bounded question an author is stuck on goes to an available peer and is settled in minutes, which is the ground of the ad hoc review (10.37g). Options still open on a draft are weighed against each other in front of the group the author has gathered, which is the ground of the informal walkthrough (10.37d). The single issue review settles a matter already posed, in a session convened for it alone, in front of reviewers who examined the product before walking in. Options are weighed along the way; what the session aims at is the decision.
Five objectives escape it and all five for the same reason, its scope. Remove defects: BABOK gives this objective to the inspection, by name, in the same list of definitions, "the focus is to remove defects and to create a high quality work product" (§10.37.3.2). What a single issue review finds is a departure from the standard in focus, and what it leaves is a decision on that departure; the defect log and the verification of the rework belong to the inspection (10.37a). Ensure the work product is complete and correct: completeness is a judgement about the whole product, and this review's subject is one matter. The whole product belongs to the inspection (10.37a) and to the formal walkthrough (10.37b). Establish consensus on an approach or a solution: an agreement on an approach also bears on the whole, and a deliverable a team has to sign off belongs to the formal walkthrough (10.37b); the agreement a single issue review produces bears on one point. Educate reviewers about the content: a session convened for one question teaches the question. The product is walked through in front of a group in a formal walkthrough (10.37b) or an informal walkthrough (10.37d), and it is put under the eyes of several independent readers in a pass around (10.37f). Measure work product quality: a measurement presupposes a defect log, and BABOK gives the logging and the consolidation to the inspection (10.37a). Wiegers draws the same gradation in his Table 3-2: data collection and analysis is marked yes for the inspection, "maybe" for the team review and no for the walkthrough.
The deliverable is the decision on the matter in focus. It names the departure, it says what changes, it says who carries it and by when. Out of it come no defect log and no consolidated set of findings: those are the artefacts of the inspection (10.37a) and of the formal walkthrough (10.37b). Nor does the session come back to check that the changes decided have landed: it settled the matter, and the point decided is then followed in item tracking (10.26).
Usage
When to use it
- One standard is at stake: BABOK's definition names this case, "either one issue or a standard" (§10.37.3.2).
- A precise question has to be settled: it is posed before the session, and the session exists for it.
- The answer needs several competences: legal, architecture, business, brought together once around the same point.
- The reviewers can examine the product alone before the session: the definition makes this the technique's condition.
- The decision has to hold in writing: it carries a departure, an owner and a date.
- One point is blocking an approval: the review is formal, and an approval needs something to rest on.
When not to use it
- The whole product has to be combed through (defects, completeness, correctness): an inspection (10.37a) logs the defects and verifies the rework; a formal walkthrough (10.37b) consolidates as a team what each reviewer found alone.
- The question is open and the document is a draft: options are weighed in an informal walkthrough (10.37d), and a bounded question goes to a peer in an ad hoc review (10.37g).
- One competent reading is enough and no session is warranted: a desk check (10.37e) puts one reviewer in front of the document and the standard; a pass around (10.37f) puts several there, each on their own.
Description
The size of the question
BABOK gives this technique no steps. What decides the outcome lies in the way the matter is posed, and that is played out before anyone walks into the room.
Pose one matter only
"Does our data model comply with the revised FADP?" is a work programme. "Does the Customer entity of the data model comply with art. 6 revFADP, purpose and proportionality, and with art. 5 let. c on sensitive data?" is a matter: it can be prepared, it can be settled and one session is enough for it. A matter fits into a question, which is what separates it from an agenda.
Name the standard by its articles
ISO/IEC 20246 describes the technical review as the one that "identifies discrepancies from specifications and standards" (§3.18), and a discrepancy is measured against a text rather than against an intention. The standard comes into the session with its article numbers and serves as the checklist during preparation.
Choose the reviewers on competence
IEEE speaks of a team of qualified personnel (§3.7), ISO of a team of technically qualified personnel (§3.18): what gets a person into the room is competence on the matter. BABOK notes, for its part, that a manager sometimes attends a review because of their expertise and that the moderator then takes care to preserve the openness of the exchange (§10.37.3.3).
Require the prior examination
BABOK's definition rests entirely on it: the reviewers perform a careful examination of the work product before the joint session. IEEE makes it an obligation for its own instrument: each member of the review team shall examine the product before the meeting, and the review leader shall verify that the members are prepared, finding a stand-in or rescheduling the session if need be (§5.5.5). A session where nobody has read starts reading, and reading in the room is a different technique.
Where the single issue review sits
BABOK lists its seven review types in descending order of formality and calls the first three formal techniques, the next four informal techniques (§10.37.3.2). Two binary questions cut across that ordering: is the session planned, does the reviewer examine the product alone before speaking? The single issue review answers yes to both, exactly like the inspection (10.37a) and the formal walkthrough (10.37b). The three formal reviews carry the same signature: on the axis, the single issue review stands on the formal, planned, prepared side, and what separates it from its two neighbours is read elsewhere. Wiegers marks those two activities for six of the seven types (Table 3-1); the single issue review, absent from his spectrum, takes its two yeses from BABOK itself, which describes it as a careful individual examination prior to a joint session.
The same name in IEEE and in ISO
BABOK itself gives this review its other name, technical review, which the two standards of the field carry. It is worth reading what they say about it, because their instrument is broader than BABOK's.
ISO/IEC 20246 defines the technical review as a formal peer review, conducted by a team of technically qualified personnel, which examines the suitability of the work product for its intended use and identifies discrepancies from specifications and standards (§3.18); it counts it among the review types it defines (§5.3). The same standard defines a formal review as one that follows a defined process and produces a formal documented output (§3.5): that is where the record comes from, and ISO says it of its own instrument. IEEE 1028 gives the same definition, of the software product (§3.7), and goes further: four roles shall be established for a technical review, decision maker, review leader, recorder and technical reviewer (§5.2).
That is where the two traditions part. The subject of IEEE's technical review is the software product as a whole, its suitability for its intended use; the subject of BABOK's is a question or a standard. IEEE's review recommends and refers up; BABOK's resolves the matter in focus. IEEE requires four roles, one of them a recorder; BABOK notes that a facilitator may be helpful for a single issue review and does not list a scribe among its applicable roles (Table 10.37.1). Two instruments carry the same name, so a team that speaks of a technical review has to state which of the two definitions it follows.
What the session produces
BABOK's three formal reviews are read in their three definitions, one sentence each, and they do not produce the same thing (§10.37.3.2).
| Inspection (10.37a) | Formal walkthrough (10.37b) | Single issue review (10.37c) | |
|---|---|---|---|
| What is under review | The work product | The work product | One issue or one standard |
| What the joint session does | It consolidates, as a team, the defects each reviewer logged alone | It consolidates, as a team, through the same activities, what each reviewer found alone | It resolves the matter in focus |
| What it leaves behind | A defect log and a follow-up to ensure the changes were made | A consolidated set of findings | A decision on the matter in focus |
Who facilitates, who records
BABOK gives each review role the techniques it applies to (Table 10.37.1). For the facilitator it lists the inspection (10.37a) and the formal walkthrough (10.37b), which it also calls a team review, and then: "may be helpful for single issue review". For the scribe it lists the inspection (10.37a) and the formal and informal walkthroughs, that is the formal walkthrough (10.37b) and the informal walkthrough (10.37d), and the single issue review is not on that list.
BABOK does not bring those two lines together. Bringing them together is worth doing all the same, because the single issue review is the one of the seven that exists to produce a decision. The facilitator, in that same table, is further charged with validating that the reviewers have examined the work product before the session begins, the very property the definition of the single issue review rests on, which describes them examining the product before the joint session. The duty the definition rests on is the one BABOK houses in the role it says only may be helpful for this review.
What to do about it is decided before the session and comes down to two appointments. Whoever convenes a single issue review names the person who will write the decision, because the decision is the deliverable and a deliverable has an author. And they name someone to validate that the reviewers have read, because the whole technique rests on that reading.
The traps
Widening the matter in session
The session was convened for one question, the reviewers prepared that question and a second one arrives on the agenda because everyone is in the room. Nobody has prepared it, and a session can settle only what has been prepared. A matter that arrives mid-session goes back into preparation or changes review.
Taking it for an inspection
It is formal, it is prepared, it holds a session, and it is easy to conclude that the product has been combed through. What was examined is only what the matter called for. The evidence that a whole product was examined defect by defect, with a verification of the rework, is manufactured in an inspection (10.37a).
Leaving the session without an owner or a date
"We agree that attribute has to go" is an agreement, and it lives three weeks. A decision carries an owner and a deadline, and it enters item tracking (10.26) the same day. That is the only thing separating a decision from a shared opinion.
Leaving the standard at the door
A session that debates what the standard means has not prepared the standard. The text, its articles and its paragraphs are on the table, and the disagreement then bears on the product rather than on the law.
AI considerations
The matter of a single issue review is a confrontation between two texts, a work product and a standard, and that is work a large language model does quickly and well. It is also the review of the family where the model's role is easiest to overrate, because the review produces a decision and a model produces an answer.
Prepare the confrontation
Running the 24 attributes of an entity against the declared purposes, flagging the ones that have none, picking out the ones that fall into a particular category of the text: it is mechanical, it is tedious, and it is exactly what a model returns in minutes. The reviewers' hour is then spent on the contested cases rather than on the sorting.
Paste the text of the standard into the context
The revised FADP came into force on 1 September 2023 and the article numbering changed. A model answering from memory often answers on the old law, with the assurance of the new one. The standard is pasted into the context, article by article, and the answer is checked against it.
The decision stays human
The session exists to take a decision that binds the organisation, and a decision binds someone. A model that writes "this attribute should be removed" has produced a recommendation, and the value of the single issue review lies in its producing something else: a point settled, by named people, on a date.
The prior examination still has to happen
A well-kept list of findings, machine-generated, makes a session where three people are discovering the subject look like a prepared one. The definition of the technique rests on the reviewers having examined the product, and nobody can validate that in their place.
Data protection
A data model that carries health data does not get pasted into a public model: that is a revised-FADP question before it is a productivity question, and the organisation's rule, which model, which data, hosted where, is settled before the session rather than during it.
Examples
What a single issue review produces is the written resolution of the matter in focus.
| Matter in focus | Does the Client entity of the data model comply with art. 6 revFADP (purpose, proportionality) and art. 5 let. c (sensitive data)? |
|---|---|
| The standard | revFADP, art. 5 let. c (sensitive data), let. f (profiling) and let. g (high-risk profiling); art. 6 para. 2, para. 3 and para. 7; art. 7 (data protection by design and by default) |
| Reviewers | The data protection adviser, the data architect, the product owner of the loyalty programme |
| Prior examination | Each held the 24 attributes of Client against the programme's declared purposes, alone |
| Finding | 3 attributes out of 24 with no declared purpose: date_naissance_conjoint, revenu_estime and etat_sante_declare, which is sensitive data within the meaning of art. 5 let. c |
| Decision of the session | etat_sante_declare and date_naissance_conjoint leave the model. revenu_estime stays: its purpose is declared. It is an inferred attribute, therefore profiling within the meaning of art. 5 let. f, which names the economic situation, and the session holds it to be high-risk profiling; express consent enters the registration path (art. 6 para. 7 let. b). The model is compliant once the three changes are made. |
| Follow-up | Carried by the data architect, due 31 August, tracked as an open item |
The session was convened for one question, and it ended with an answer: three attributes, three decisions, with an owner and a date. Two of them take a piece of data out of the model. The third keeps one and adds to the registration path the express consent that art. 6 para. 7 let. b revFADP requires for high-risk profiling by a private person, which the retailer is. The judgement passed on the risk of that profiling is the very matter the session had to settle.
The three reviewers had walked in prepared, and one session was enough. The review does not come back to check that the three changes landed in the model: it settled the matter, and the point stays open until it is closed, followed in item tracking (10.26).
Visualizations
The formality spectrum is a spatial object: seven types placed on an axis, two marks per cell. It locates without distinguishing: the three formal reviews answer yes to both questions. The figure teaches the position of the single issue review on the formal, planned, prepared side. An axis and positions do not render as cells, so it is drawn.
What separates the single issue review from its two neighbours is its scope, and a scope is read by comparing three definitions term by term. What is made of rows and columns renders as cells. The three columns come from one source, BABOK's §10.37.3.2, and the table's caption says so where it has to be said, in the table.
The artefact produced is the minutes of a single issue review, with the matter, the standard, the finding and the decision. The choice between the seven reviews belongs to the parent technique (10.37), which exists for that.
Cost
| Phase | Level | Rationale |
|---|---|---|
| Preparation | Medium | The matter is posed and the standard named before the session, and each reviewer examines the work product alone: BABOK writes this into the very definition of the technique (§10.37.3.2). The preparation is real and it is paid for, but it bears on one question rather than on a whole product. |
| Execution | Medium | A formal session, several people in the same room at the same time, one matter on the agenda. The cost is that of the session multiplied by the number of reviewers, and it is bounded by the scope. |
| Documentation | Medium | The decision is the deliverable: it has to be written, it names a departure, an owner and a deadline, and the open point enters item tracking (10.26). BABOK does not list a scribe among the roles applicable to this review (Table 10.37.1), so someone is named to write it and that person's time is paid for. |
Tooling
A single issue review needs three things: the text of the standard with its article numbers, the work product in a form each reviewer can annotate alone and somewhere for the decision to live after the session. The first is a document, the second is the tool the product was built in, the third is item tracking (10.26).
The standard serves as the checklist during preparation, article by article, against the product. A mapping table between the articles of the text and the elements of the product, kept in a spreadsheet or in the requirements management tool, is what makes a one-hour session possible: every reviewer arrives with the same grid, and the session bears on the cells where the three readings diverge.
Review platforms and comment threads are the natural tool of the individual examination, each reviewer on their own. The session, for its part, is held: it exists to settle the matter, and a decision is taken where the people are. The day the exchange happens entirely in writing and no session is held, it is a pass around (10.37f), and BABOK notes that comments exchanged by email can produce a great many messages to work through, which makes it difficult for the author to resolve the disagreements (§10.37.4.2).
Sources
- IIBA, A Guide to the Business Analysis Body of Knowledge (BABOK Guide) v3, §10.37 Reviews (§10.37.3.1, the seven objectives; §10.37.3.2, the seven types; §10.37.3.3, the participants; Table 10.37.1, the roles; §10.37.4.2, the limitations).
- Karl E. Wiegers, Peer Reviews in Software: A Practical Guide, Addison-Wesley, 2002, chapter 3, "Peer Review Formality Spectrum" (Table 3-1, activities by review type; his spectrum counts six methods and does not include the technical review).
- ISO/IEC, ISO/IEC 20246:2017, Software and systems engineering: Work product reviews (§3.5, the formal review; §3.18, the technical review; §5.3, the review types).
- IEEE, IEEE Std 1028-2008, IEEE Standard for Software Reviews and Audits (§3.7, the technical review; §5.2, the roles; §5.5.5, the preparation).