Skip to content

CHANGELOG.md

Latest commit

 

History

History
74 lines (56 loc) · 9.57 KB

CHANGELOG.md

File metadata and controls

74 lines (56 loc) · 9.57 KB

Change Log

All notable changes to this project will be documented in this file. This change log follows the conventions of keepachangelog.com.

[Unreleased]

Changed

  • IMPORTANT: formform versions ≥ 0.4.0 are now distributed under Mozilla Public License 2.0, changing from the previous MIT license. This is because I want to ensure that any changes/improvements made by other people to the source files will be shared with everyone in the spirit of free and open source software. Still, I don’t want to impose any such restrictions on people who just want to use formform in their projects as a library, hence a weak copyleft license.

[0.3.0] - 2025-12-23

Added

  • When specifying a position in figure-related ini specs, users can now also provide a ratio (a float number in the unit interval as a fraction of the size of a generation) instead of an index number.

Changed

  • emul/ca-iterator and emul/create-ca now take an options map where users can set the seed for random inis. In the case of emul/create-ca, the history-cache-limit must now be set in this options map instead of as an extra argument.
  • Ini type :figure-repeat now takes a figure ini (type :figure or :rand-figure) instead of a background ini and a pattern.

[0.2.0] - 2025-11-07

Added

  • calc/vmap-dimension to calculate the dimension of vmaps (like dna-dimension) or get it from meta :dim cached by a vmap constructor (not for zero-dimension vmaps)
  • expr/permute-vars to generate all permutations of a variable ordering
  • expr/formDNA-perspectives as a complement to calc/dna-perspectives which does the same thing but while transforming a formDNA expressions into an arrangement of formDNA expressions as its perspectives (keeping variable names)
  • io/print-dna and io/print-const functions such that formDNA and constants not wrapped in expressions can be printed properly in formula notation
  • calc/dna-get as a convenience function to quickly get a specific value from a dna (also because it isn’t straightforward that filter-dna can do the same thing with a little more effort)
  • calc/vmap-perspectives to complement calc/dna-perspectives (which it takes as input)
  • New operator :tsds provides a syntactic shortcut to build triple selective decision systems with binary selections. The 3 terms can be arbitrary expressions. Its FORMula notation looks like this: [:tsds 100100 L,E,R]
  • New formform.emul module to (declaratively) represent, compose and calculate cellular automata with FORM logic. The API is mostly complete, but small changes may still happen. An introduction and more detailled explanation is in progress.
  • nuim-code and nmui-code definitions in the API for convenience
  • rand-const, rand-dna and rand-vpoint can now all take a seed for reproducability in randomization
  • Added rand-const-weighted, rand-dna-weighted and rand-vpoint-weighted variants, which take a weights argument for uneven probability in random constant generation
  • eval->val/==> and eval->val-all/==>* are two new functions in the expr module. The former only returns simple values, including “holes” :_. The latter returns only complex values, i.e. raw formDNA (possibly with holes). In essence, they return output that can be processed by functions in the calc as well as the new emul module.
  • eval-tsds->val-all behaves just like ==>*, but takes a 6-digit binary selection for a TsDS expr., which is especially convenient for emul functions
  • Added calc/partial-dna? predicate to check if formDNA contains holes.
  • Extended calc/rel and calc/inv to enable correct arithmetic with value holes.
  • calc/equal-partial-dna? and calc/equiv-partial-dna?, which work exactly the same as their non-partial counterparts, but allow value-holes in the formDNA input, where they are assumed to be equal (:_ = :_ is always true).

Changed

  • Changed constant/formDNA to lowercase in FORMula notation as well as in formform syntax. It is now much more pleasent to read than with uppercase and more aligned with the conventions in mathematical texts such as uFORM iFORM and LoF, which was the primary reason for this breaking change. Please also read about the change to formDNA notation below that went along with this one.
  • Reversed the order of formDNA, which now reads left-to-right instead of right-to-left. The main reason for this change is to make formDNA more intuitive in reading and writing and reduce the mental load to reverse the order in your head. I apologize for this breaking change if you were already using formDNA; there will most certainly be no further changes to the notation. It went along with the new lowercase convention for constants, so that your existing formDNA data will not suddenly become incorrect and instead throws an exception, so you can safely make the transition.
  • Removed 0-arity in rand-vpoint because the new random functions don’t return an infinite lazy seq (I believe that has never been useful anyway).
  • rand-dna doesn’t take an elems vector anymore to select from, which could have arbitrary elements. Its primary function is now better served by the -weighted variant and this ensures that it always returns a valid formDNA.
  • expr/mark-exprs is deprecated to be replaced by expr/form-marked and expr/make-marked (the “unmarked” variant), to avoid the need for an options map.
  • expr/nest-exprs is deprecated to be replaced by expr/form-nested-l, expr/form-nested-r (l/r distinguishes the nesting direction) and expr/make-nested-l, expr/make-nested-r (the “unmarked” variants), again to avoid the need for an options map.
  • expr/simplify-expr-chain is deprecated to be replaced by expr/simplify-nested-l and expr/simplify-nested-r (l/r distinguishes, again, the nesting direction). This is in line with the change above.
  • expr/eval->expr-all (alias expr/=>*), expr/eval-all, calc/permute-dna, calc/dna-perspectives, calc/vdict and calc/dna->vdict now take opts as their last argument instead of the first one. This is to make argument order of the options map consistent with other functions in formform and with the convention in the Clojure community.
  • expr/eval->expr / expr/=> now returns the simplified expression instead of a hole :_ for uninterpretable input (such as unregistered symbols or uninterpreted variables). Likewise, expr/eval->expr-all / expr/=>* returns the pre-simplified, uninterpreted expression if it could not be determined to a value in any of its interpretations. It will also just return simple values instead of wrapping them in a formDNA expr. There are some new options explained in the docstring.
  • evaluate now separates output values in :result and the simplified expression in :simplified. The result is now nil if it could not be determined to a value. For eval-all, results are either value constants or nil, but can optionally be maps with the same entries as from evaluate. Options are explained in the docstring.
  • Renamed calc/equal-dna, calc/equiv-dna, expr/equal and expr/equiv to have ? appended to their name, to follow the convention for predicates.
  • Changed calc/reduce-dna-seq to check if the dna-seq contains value holes (:_) and in that case just return it, because reduction with value holes is not possible (extension is, though!).
  • expr/simplify and expr/simplify-in now treat each instance of an expression/value hole (:_) in an expression/context as a unique, distinct symbol. This means that holes in subexpressions do not get simplified by law of calling like they used to.
  • calc/reduce-dna-seq now takes the optional terms as its second argument instead of the first. This is part of an effort to make argument order more consistent across formform.
  • calc/reorder-dna-seq, calc/expand-dna-seq and calc/reduce-dna-seq have been renamed to calc/reorder-dna, calc/expand-dna and calc/reduce-dna, because the -seq suffix doesn’t make sense anymore and would only confuse users. Aliases to the old names are kept for now but might be removed later.
  • expr/interpret, expr/interpret*, expr/interpret-walk and expr/interpret-walk* now all receive the optional env as the second argument instead of the first. This makes the api more consistent with the other functions in expr that receive an env.
  • Removed expr/simplify-op as well as expr/simplify-sym, since they can both be (almost) fully replaced by expr/simplify, which is less confusing. The only difference is that if the final output is a constant, it will be further simplified.

Fixed

  • Macros not recognized in ClojureScript
  • calc/permute-dna-seq, which had an incorrect implementation and lack of testing
  • calc/filter-dna to handle the edge-case where formDNA dimension is 0, i.e. the dna is equivalent to a constant
  • calc/dna-perspectives now retains its original key-order in metadata after being converted to a map, which is useful if the order in which perspectives are listed matters
  • expr.core/simplify-content used to call the method interpret-sym for expression symbols, which was incorrect, it now calls simplify-sym

[0.1.0] - 2023-05-07

First alpha release of the rewrite on Clojars. I advice against using it in any serious manner just yet, as there are still some rough edges, missing tests, unresolved issues, undecided questions and maybe also some outdated documentation.

However, you are welcome to experiment with the library (take a look at the introduction first) and I am thankful for feedback of any kind, as this is my first real Clojure project.