What people remember about a spreadsheet is rarely the length of the feature list; it is whether opening and calculating feel crisp.

A spreadsheet that came out of the GNOME desktop

Gnumeric is an open source spreadsheet maintained by the GNOME community. The official site compresses its own position into three short headings, placed side by side on the first screen: free, fast to start, and accurate in calculation. Three lines, and a reader can see at once where the project spends its effort.

Its origin is written into the first line of the repository description: the spreadsheet of GNOME. The sentence that follows states the aim, which is to become a replacement for commercial spreadsheet software. The wording stays modest, and the emphasis lands on two things: existing files must open, and formulas must calculate correctly. Setting up a separate format and asking people to relearn is explicitly not the plan. The project has walked that road for many years, and the commits and tags accumulated in the repository are substantial by any measure.

The licence is stated just as plainly. The file in the root directory says a user may pick either version 2 or version 3 of the GPL, and that dual arrangement removes the need for a separate negotiation whenever the version question comes up.

Starting quickly and counting straight come first

Two complaints about spreadsheet software are old. One is waiting for a large sheet to open. The other is a function that returns something other than what the reader expected. Putting both on the front page amounts to publishing the numbers the project cares about. The impression in use matches the claim: the program itself is not heavy, the start is crisp, and a large sheet does not turn sticky when scrolled.

The function library is where most of the investment has gone. Beyond the usual mathematical, statistical, date and text functions, version 1.12.62 adds statistical entries such as PHI, GAUSS and PERCENTOF, and three regular expression functions: REGEXTEST to decide whether a string matches, REGEXEXTRACT and REGEXREPLACE to pull text out and to substitute it. Date functions that take arguments, NETWORKDAYS.INTL and WORKDAY.INTL among them, were filled in during the same round, which saves a good deal of manual work when counting working days. For bulk cleaning of text, a regular expression is usually less trouble than a stack of nested find and replace steps.

What changed in 1.12.62

The new release arrived in September 2026. The change list is mostly functions and compatibility: several new functions, improvements to xlsx export, adjusted behaviour in a number of existing functions, and a correction to the precision of GEOMEAN. The list also mentions a batch of fixes for problems that affected Wayland users, and for anyone working on a Linux desktop day to day those repairs are easier to notice than new features. The precision fix carries the same kind of value: the same set of numbers now lands closer to the mathematical definition, and the cross-checking that used to be needed goes away.

There is nothing fancy about the way it is obtained. The official advice for ordinary users is to go through the distribution repositories and let the system install the dependencies in its standard manner. Anyone willing to compile can take the source from the GNOME code repository, where the version history and commit record are open to inspection. That distribution model also fixes the audience: mostly people already settled on a Linux desktop, rather than someone looking for a stopgap in a hurry.

The same point marks where to weigh it up. Distribution runs through the distributions and the source, the desktop build is aimed mainly at Linux, and while file interchange is handled well, with the common formats read and written, the niche it occupies suits people who work on Linux for the long run and want their spreadsheet tool a little lighter.

Light is not a missing feature; it is leaving the resources to the step that actually calculates.

A practical note: if the day is spent on a Linux desktop and the work involves medium sized numeric and statistical tables, install it straight from the distribution repository; function naming follows the common spreadsheet conventions closely, so the migration cost sits mostly in shortcuts and interface habits.