How did computer programming categories change?

Computer programming categories in ISIC Rev. 5, UN Statistics Division, sit within new Section K, which covers telecommunications, computer programming, consultancy, computing infrastructure and other information services. That placement reflects the split logic: programming is grouped with the computing and information service infrastructure, not with publishing and content production in new Section J. Within Section K, the divisions, groups and classes, and their explanatory notes, give the detail, and a reader should check them before using any specific code.

Why does programming sit with infrastructure rather than content in Rev. 5?

The opening paragraph gives the placement, and the reason for it is worth pausing over. Section K gathers the machinery and services that make information usable: the networks that carry it, the programming and consultancy that build and support systems, the computing infrastructure that runs them, and the other information services around them. Section J, by contrast, gathers the making and distribution of content itself. A programming team and a publishing house may both work with text on screens all day, and their offices may look alike, yet the Rev. 5 grouping separates them because what each produces for counting purposes is different in kind. One produces a working system or a service around a system. The other produces material intended to be read, watched, or listened to. That distinction is the whole split in miniature, and it is why a reader who starts from the section wording, rather than from a familiar job title, is less likely to drift into the neighbouring section.

What changed for a reader who learned the old single section?

Under the old arrangement, a reader could stay inside one broad information block and still feel oriented. Rev. 5 removes that comfort, deliberately. The reader must now decide, at section level, whether the activity in front of them belongs with content or with the infrastructure around information, and only then move down to divisions, groups, and classes. That extra decision is not bureaucratic decoration. It changes which totals the activity contributes to, which neighbouring categories sit beside it in a table, and which explanatory notes must be read before a figure is quoted. A time series that crosses the revision therefore needs a note at the point of change, because a programming total compiled under the old block and a programming total compiled under Section K are not automatically the same population, even when the words in their titles overlap. The honest move is to name the revision beside every figure, and to treat any comparison that mixes revisions without a note as unfinished work.

How does a survey team keep programming figures steady across the revision?

Picture a survey team revising its questionnaire after adoption of Rev. 5. The old coding manual pointed coders to a single information section. The new manual must point them first to the J or K choice, then to the division and class notes inside Section K for programming work. If the questionnaire, the manual, and the published table are updated at different speeds, coders will answer from one version while readers interpret from another, and the drift will show up as a sudden, unexplained movement in the published series. The team avoids that by updating the three documents together, by testing the new instructions on a batch of real responses, and by keeping the version label, ISIC Rev. 5, UN Statistics Division, visible at each stage, from the first coding decision to the final chart. Where a response is too thin to justify a fine class, the team reports at the broader level that the evidence does support, and records the gap, rather than inventing precision that a later checker cannot reproduce.

What should a reader carry away before quoting a programming figure?

Read the division and class notes inside Section K before settling on any detail, and check the excludes lists with particular care, because neighbouring titles in computing and information services can look close while pointing to different scopes. Ask what level the figure sits at, because a section total and a class figure answer different questions even when they share a word. Ask what period the figure covers, because national adoption dates can fall inside a series and create a break that has nothing to do with real change in programming work. And keep the wider system in view: a programming figure feeds business registers, national accounts, and related tables, so an error at the coding stage travels further than the single table where it first appears. That is why revisions are documented and phased in, rather than slipped in silently.

This guide covers statistical classification only, and it leaves every individual determination open: it does not settle any reader’s code, principal activity, or registration, and any code for a specific record should be confirmed with the national statistical institute or business register that holds the record.