Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

What is the rationale for making this project AGPL? #551

Closed
jlumbroso opened this issue Apr 2, 2021 · 13 comments
Closed

What is the rationale for making this project AGPL? #551

jlumbroso opened this issue Apr 2, 2021 · 13 comments
Assignees

Comments

@jlumbroso
Copy link

I noticed the initial license for this project was Apache, but early in the commit logs this was replaced by the AGPL (I think by @tarheel at fe95903).

I am an RCV activist as well, and I am sincerely confused by the approach.

Strong copyleft licenses are activism for open-source. Here it seems like this project is both advocating for RCV but also for strong open-source, and that seems like a mix of concerns.

If your goal is to ensure that RCV comes to the US, wouldn't you welcome private vendors (who cannot be cut out of the process since a large number of election-related contracts are private) having access to this software and being able to use it themselves — this would not be great for open-source, but it would be great for RCV, as it would reduce the friction necessary for private actors to move to RCV.

I can understand someone considering that Apache is too permissive, that's why my favorite license is LGPLv3 — both it and the Mozilla Public License v2 are good trade-offs between promoting open-source but not making it impossible for private companies to use the artefacts. It states that if a vendor uses your library, they can do so without making their work open-source; but if they make any changes to your library, they must make these publicly accessible. I believe this is a good compromise between encouraging adoption and protecting the rights of the codebase.

@tarheel
Copy link
Contributor

tarheel commented Apr 7, 2021

Hi @jlumbroso, thanks for the question! You make good points. Our original decision to adopt AGPL was in response to strong encouragement from open-source proponents, as you may have guessed. You're right that a license like LGPL would make it much easier for a vendor to incorporate our code into their products. In practice, though, we've found that it's unlikely that a vendor would want to engage with our work in this way; they're more likely to a) merely look to our code as inspiration for their own proprietary implementation, or b) keep their focus on the voting machines and cast vote record generation and let our software handle the RCV tabulation directly (as is now happening in multiple jurisdictions). If we did encounter a situation in which a vendor wanted to pull our work into their own codebase and we saw potentially significant benefit for the broader RCV movement from enabling that adoption, we would give serious consideration to modifying our license. But, as far as we're aware, this is only a hypothetical at the moment.

@HEdingfield
Copy link
Contributor

Since we're going to be modifying the license in the near future (CC: @chughes297, @moldover), let's also make sure to update all the copyright blocks to say 2022 (they currently say 2020) and fix this typo where it says "Universal RCTab" instead of "RCTab".

@moldover moldover self-assigned this Jul 11, 2022
@moldover
Copy link
Contributor

This is from an email from @chughes297 , just to keep @jlumbroso and everyone up to date:

Hi Hylton, Jon, Louis, and Tamir,

Hope you're all doing well.

This email is a brief update on our licensing strategy for RCTab. TL;DR: we are moving RCTab entirely over to an MPL 2.0 license and dropping AGPL entirely.

Rene and I spoke with Heather and Greg at OSET and they advised us that the most realistic path forward is to offer RCTab under a single license - namely, MPL 2.0.

I assumed it was our best strategy to offer RCTab under AGPL via GitHub/to non-commercial licensees and under MPL 2.0 for commercial licensees. Heather explained, though, that once you provide the software under MPL to one person you need to provide that same software version under MPL to all potential users.

I also thought keeping the GitHub version under an AGPL license would give us more bargaining power with the vendors. Both Greg and Heather thought this wouldn't really pan out - because one vendor (Hart) has already accepted RCTab under the MPL, that sets the stage for other vendors to accept the same license. We may need to be ready to educate the vendors to get them on board with MPL, but both Greg and Heather assured me we wouldn't lose anything in a negotiation with the vendors if we're already providing the software under MPL 2.0. It will also be administratively easier to just provide a single license and not track which users are using the software under which license.

To sum up: we're just going to provide RCTab under a single license - MPL 2.0. We are dropping AGPL.

One aside that I want to note:

MPL 2.0 does also allow you to license derivative works under AGPL or other more-copyleft licenses, too, so we could still provide versions of RCTab under more aggressive licenses (if we want to).

Two things I need you guys to do to officially move us over to MPL 2.0:

  1. Run a BLAME function on the RCTab code to double-check that we've captured all contributors to the software on the license wavier
  2. Update the license on GitHub to MPL 2.0 (from here: https://www.mozilla.org/en-US/MPL/2.0/).

Let me know if you have any questions.

Thanks!

@moldover
Copy link
Contributor

moldover commented Jul 11, 2022

So @chughes297 @rrojas350 we have the the current copyright header text at the top of every source code file, and we need to know what the new text should be:

RCTab
Copyright (c) 2017-2020 Bright Spots Developers.
This program is free software: you can redistribute it and/or modify it under the terms of the
GNU Affero General Public License as published by the Free Software Foundation, either version 3
of the License, or (at your option) any later version.
This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY;
without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See
the GNU Affero General Public License for more details.
You should have received a copy of the GNU Affero General Public License along with this
program. If not, see http://www.gnu.org/licenses/.

@jlumbroso
Copy link
Author

@moldover @HEdingfield I believe this is an excellent decision, and in fact one that might make a difference for the better!!

Well done all of you! 👏🏻👏🏻👏🏻👏🏻

@chughes297 Let me know if you would like help with the headers.

@rrojas350
Copy link

@moldover I believe an easy fix would be to replace GNU Affero General Public License with Mozilla Public License V2.0, just find and replace. Something like this:

RCTab
Copyright (c) 2017-2020 Bright Spots Developers.
This program is free software: you can redistribute it and/or modify it under the terms of the
Mozilla Public License Version 2.0 as published by the Mozilla Foundation.
This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY;
without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See
the Mozilla Public License Version 2.0 for more details.
You should have received a copy of the Mozilla Public License Version 2.0 along with this
program. If not, see https://www.mozilla.org/en-US/MPL/2.0/.

@chughes297 will have to chime in if that still covers us from a legal standpoint.

@moldover
Copy link
Contributor

@rrojas350 the current copyright text is specific to GNU: https://www.gnu.org/licenses/gpl-howto.html
An ad-hoc find/replace does not seem like a good idea to me.

Can you please follow up with @chughes297 IRL and make sure he is looped into Github conversations, and find out what, if any, new copyright text should go into the source header files?

@tarheel
Copy link
Contributor

tarheel commented Jul 15, 2022

According to his email auto-response, Chris was out of the office for a while but should be back now.

@rrojas350
Copy link

Maybe I should preface my comments with, "this does not constitute legal advice."

Chris just got back today and is catching up on emails, so I'm sure he'll chime in soon. I'll shoot him a text just to verify.

@moldover
Copy link
Contributor

I found these instructions: https://www.mozilla.org/en-US/MPL/headers/
They have a similar setup and go into the source code headers so it could look like this:

/*

  • RCTab
  • Copyright (c) 2017-2022 Bright Spots Developers.
  • This Source Code Form is subject to the terms of the Mozilla Public
  • License, v. 2.0. If a copy of the MPL was not distributed with this
  • file, You can obtain one at https://mozilla.org/MPL/2.0/.

*/

@chughes297
Copy link
Collaborator

Sorry for my delay here - was closing out a few other projects earlier this week! This looks right to me.

moldover added a commit that referenced this issue Jul 21, 2022
@moldover
Copy link
Contributor

No worries! I got the sources all update I think so it should be gg.

moldover added a commit that referenced this issue Jul 21, 2022
* new license file

* new contributors guide

* cleanup from rebranding

* Update contributor_guide.md

* Update contributor_guide.md

* update license to MPL 2.0 and date to 2022 in source headers, test_data, intellij copyright template #551
@moldover
Copy link
Contributor

closed via #604

moldover added a commit that referenced this issue Aug 12, 2022
moldover added a commit that referenced this issue Aug 15, 2022
* Changed package from com.rcv to network.brightspots.rcv.

* remove unused tests: fix for #197

* add tiebreak test #197

* Allows user to cancel during interactive tiebreaker (fixes #132); adds example config file for testing interactive tiebreaker with sequential multi-seat enabled; adds commented command in build.gradle for executing CLI arguments.

* add missing precinct test

* Reconciles branch with `develop`.

* Re-adds interactive tiebreaker sample input.

* Continues fixing recently broken tests.

* Reverts name back to "Universal RCV Tabulator".

* add new random seed config field

* updated tests

* incorporate new option into tabulation logic and add test case

* onlyOneWinnerPerRound -> allowOnlyOneWinnerPerRound

* update test case to include a tie-break

* update tests

* address comments

* tests

* add test case

* update one test for each winnerElectionMode option

* updating tests

* fix tests

* fix minor output bug and add test case for previousRoundCountsThenRandom tie-break option

* fix minor issue with transfer tallies

* update test configs

* update a ton of JSON files!

* fix bad source path

* Increment version from 0.1.0 to 0.1.1. (#373)

* Increment version from 0.1.1 to 0.1.2

* Increment version from 0.1.2 to 1.0.0 (#384)

* v1.0.0

* Increment version from 1.0.0 to 1.0.1 (#389)

* Fix broken tests (#393)

* Increments version from "1.0.1" to "1.0.1_20200320_internal" (#423)

* Adds multiSeatBottomsUpPercentageThreshold (#428)

* Tests for Dominion CSV conversion routine (#429)

* test for Dominion csv conversion routine #415

* 2 more tests for cvr conversion

* updates for Dominion conversion tests: PR #425

* fix for PR #425

Co-authored-by: jonmoldover <jonmoldover@gmail.com>

* Adds AUTOMATED_TEST_VERSION code so tests can remain stable as app version updates (fixes #426). (#435)

* Fixes Dominion conversion output filenames and path (#436)

* Dominion conversion output files now output to same directory as input files and include timestamp in filename (#408).

* Updates Dominion tests to account for new filename schema; Dominion test files are now cleaned up automatically; fixes bug with fileCompare not closing readers.

* All delete operations in `TabulatorTests` now handle exceptions with useful error messages if they fail.

* read and write precinct portion and precinct (#433)

* read and write precinct portion + precinct

* edits for precinct portion PR#433

* Update src/main/java/network/brightspots/rcv/DominionCvrReader.java

Co-Authored-By: HEdingfield <HEdingfield@users.noreply.github.com>

* Update src/main/java/network/brightspots/rcv/DominionCvrReader.java

Co-Authored-By: HEdingfield <HEdingfield@users.noreply.github.com>

Co-authored-by: HEdingfield <HEdingfield@users.noreply.github.com>

* UWI test (#440)

* add UWI test files

* add an additional test to validate my fix of #397

* use a better name for the first test

* make batch elimination work properly with singleSeatContinueUntilTwoCandidatesRemain (#441)

* make batch elimination work properly with singleSeatContinueUntilTwoCandidatesRemain

* cleaned up wording

* v1.1.0 (#454)

### New features:
* Added support for converting Dominion JSON CVRs to generic .csv format (including precinct portions) (#404, #406, #407, #408, #415, #439)
* Added multiSeatBottomsUpPercentageThreshold option (#403)
* Added CLI option to convert Dominion CVR to generic .csv (#408)
* New GUI menu and conversion options (can now convert to CDF and convert Dominion to generic via the GUI) (#408, #421)
* Added Dominion Alaska CVR to sample_input folder

### Bug fixes:
* Batch elimination now works properly with singleSeatContinueUntilTwoCandidatesRemain (#396)
* In a multi-seat contest, if someone wins in the first round, we now automatically eliminate undeclared write-ins before we eliminate any other candidates; previously, we treated UWIs like a normal candidate, which meant we potentially eliminated other candidates with lower tallies first (#397)
* If UWI exceeds the winning threshold in the initial count, we no longer mistakenly elect this candidate (#398)

### Backend updates:
* Updated dependencies to latest version: JDK, JavaFX, Checkstyle google_checks.xml, Checkstyle plugin, org.openjfx.javafxplugin, org.beryx.jlink, org.apache.commons:commons-csv, org.apache.poi:poi-ooxml, com.fasterxml.jackson.core:jackson-*
* Added special code to test configs to obviate the need to update the version with each increment (#426)
* Updated tests and improved test coverage
* Copyright update (#414)
* Code cleanup

* Hart (#458)

* add contest id to config

* add xml parsing module

* First cut at Hart CVR reader + test assets.

* cleanup paths and add expected results

* add full cvr data

* refactor contestId from Integer to String

* add school board test

* fix xml parsing

* logging

* re-add accidentally removed import

* cleanup for PR#548

* updates for PR #458

* updates for PR #458

* test other file reading options

* more cleanup for PR #458

* more cleanup for PR #458

* Fixes tests not working in Windows due to mangled UTF-8 BOM marker.

* Handles potential NPE if chosen directory contains no files.

Co-authored-by: HEdingfield <hylton@groupagree.com>

* Update dependencies (#459)

* Updates to latest version:
    * Gradle 6.5.1
    * JDK 14.0.1
    * JavaFX 14.0.1
    * Checkstyle google_checks.xml 8.34
    * Checkstyle plugin 8.34
    * org.openjfx.javafxplugin 0.0.9
    * org.beryx.jlink 2.20.0
    * com.fasterxml.jackson.core:jackson-* 2.11.1
    * org.junit.jupiter.junit-jupiter-* 5.6.2
* Gets tests working with Gradle 6.5.1 (fixes #283)
* Updates `APP_VERSION` to temporary value 1.2.0_alpha
* Addresses Checkstyle warnings
* Fixes versions in Hart tests

* make contestId a real config field (#463)

* make contestId a real config field

* update .txt

* Providers dropdown (#462)

* Changes "Provider" field for CVR source to be required, bases it on an enum of valid values, and changes it to a ChoiceBox at the start of the "CVR Files" tab in the GUI  (progress on #460 and #461).
* Adds log message to help identify which CVR source validation errors pertain to.
* Gets rid of `SimpleObjectProperty` warnings during compile (fixes #280).

* Hart GUI (#466)

* Provider ComboBox now dynamically enables and disables fields and buttons when different providers are selected. (Progress on #461... actually progress on GUI phase 2!)
* Enables `DirectoryChooser` for Hart so it now functions through the GUI. (Fixes #460)
* Validation improved to fail if `firstVoteColumnIndex`, `firstVoteRowIndex`, `idColumnIndex`, `precinctColumnIndex` are defined for CVRs from non-ES&S providers.
* Removes superfluous logging.

* direct Dominion tabulation (#470)

* direct Dominion tabulation

* enable specifying a Dominion CVR source path in the UI

* output generic CSV and add test coverage

* remove extra log line

* filter out CVRs from other contests during the loading

* Changes `contestId` so it's specified on a per-CVR basis (fixes #456) (#474)

* First cut at Clear Ballot rcv reader (#475)

* First cut at Clear Ballot rcv reader
Add Kansas Dem Primary regression test
Add Clear Ballot reference files
#400

* updates for PR #475

* updates for PR #475

* updates for PR #475

* Enable Clear Ballot tabulation in GUI (progress on #400) (#480)

* GUI redesign phase 1 (#492)

* Splits Output tab into new Contest Info and Output tabs.
* Redesigns GUI CVR Files tab, adds Clear button, and changes Add button so it only clears the file path to make it easier to manually enter multiple sources that share fields.
* Improves visual presentation of Candidate tab; adds Clear button and adds checkBoxCandidateExcluded when adding a candidate.
* Reorganizes presentation of rules in "Winning Rules" and "Voter Error Rules" tabs.
* Winner Election Mode and Tiebreak Mode now start undefined with all relevant fields disabled; choosing specific modes enables applicable fields.
* Changes Winner Election Modes and Tiebreak Modes to be more user-friendly, including necessary migration logic to update older config files.
* Adds `continueUntilTwoCandidatesRemain` boolean (fixes #481).
* Expands footprint of GUI window to 1200x1000.
* Implements bordered boxes.

* remove Dominion conversion feature (#487)

* update test files

* remove conversion feature from code

* update test configs

* remove unnecessary check

* allow overvote delimiter for ES&S source files (#503)

* test

* update winner election mode logic

* comments

* final comments

* Read and Tabulate Unisyn Xml Cdf (#504)

* plumbing for xml cdf reading

* add contestId to json and xml CDF parse logic
update unisyn cvr test to use the full (but still incorrect) cvr data.  This test is expected to fail.
add candidate parsing code to xml CDF parser

* dont parse Candidate data from CDF at runtime - this enforces the same constraints on validation as all other file types.  The un-used candidate parsing code is left in for implementing future helper functions which will automatically populate config files.

* Handle ContestSelections in XML correctly

* add NIST example 2 (seems to be correctly formed)

* fill out CDF XML class definitions needed for tabulation

* add some helpers

* cleanup pre-processing code

* add more class fields around contest selections
fix ranking parsing
add example Unisyn contests

* handle write-ins

* get first 2 Unisyn regression tests working :)

* fix first NIST CDF XML test

* cleanup test names

* parse GpUnit from CVRs

* remove un-used assets

* updates for PR #504

* updated data from Unisyn with GpUnit (precinct) parsing validated

* add more Unisyn regression tests
tests share a single cvr input file

* tabulate all elections - in practice we should not see more than one

* updates for PR #504

* updates for PR #504

* Sundry UX fixes (#509)

* Converts `overvoteRule` from a `ChoiceBox` to an array of `RadioButton`s; changes `overvoteRule` string display in config files and adds migration logic.
* Disables `decimalPlacesForVoteArithmetic` and `nonIntegerWinningThreshold` except when `winnerElectionMode` is "Multi-winner allow only one winner per round" or "Multi-winner allow multiple winners per round" (fixes #500).
* Adds suggested values for `overvoteLabel`, `undervoteLabel`, and ES&S column and row indices as requested by @chughes297 (overriding what was in the original GUI mock-ups).
* Clarifies language to address confusing UX in `passesBasicCvrSourceValidation()` for CVR-specific errors.

* Cdf Json Fixes (#506)

* plumbing for xml cdf reading

* add contestId to json and xml CDF parse logic
update unisyn cvr test to use the full (but still incorrect) cvr data.  This test is expected to fail.
add candidate parsing code to xml CDF parser

* dont parse Candidate data from CDF at runtime - this enforces the same constraints on validation as all other file types.  The un-used candidate parsing code is left in for implementing future helper functions which will automatically populate config files.

* Handle ContestSelections in XML correctly

* add NIST example 2 (seems to be correctly formed)

* fill out CDF XML class definitions needed for tabulation

* add some helpers

* cleanup pre-processing code

* add more class fields around contest selections
fix ranking parsing
add example Unisyn contests

* handle write-ins

* get first 2 Unisyn regression tests working :)

* fix first NIST CDF XML test

* cleanup test names

* parse GpUnit from CVRs

* remove un-used assets

* updates for PR #504

* updated data from Unisyn with GpUnit (precinct) parsing validated

* add more Unisyn regression tests
tests share a single cvr input file

* tabulate all elections - in practice we should not see more than one

* updates for PR #504

* better logging

* first cut at fixing JSON CDF reader:
- all static election objects are parsed at beginning of cvr processing
- during processing contest selections are linked to candidate objects
- candidates are stored in config - not pulled from CDF file
- fixed NIST test set 0 config and CDF json to match the NIST 1500-103 UML
- verified tabulation results are correct for Nist test set 0 test

* update ResultsWriter CVR generation code to create Candidate objects and link to them from ContestSelection objects.
verified updated NIST test 0 cvr output

* updates for PR #504

* fix typo

* update test asset with contest name

* better handling for cdf reader parse errors

* update all CDF json assets

* Updates to CDF parsing logic:
Candidate validation and CandidateId validation are done while parsing CVRs (not during static parsing)
xml parser throws unrecognizedCandidatesException (instead of just CVRParserException)
Add json getCvrContest helper
json parser logic upgrades:
 handle writeIns in both places
 check for rank in both places
 check for HasIndication
 check for IsAllocable
 synthesize cvrId

* cleanup for PR #506

* updates for PR #505

* Throw if undeclared write-in is found but label has not been defined.

* make some fields per-source instead of per-config (#517)

* create new class

* helpful errors in CLI; don't allow newer version in config than in app

* make fields per-source

* comments

* Jon's changes

* fix almost all tests

* fix Dominion tests

* update validation and documentation for 4 fields that moved to CVR level (#518)

* update validation and documentation for 4 fields that moved to CVR level

* Fixes broken tests and a couple of typos.

Co-authored-by: HEdingfield <hylton@groupagree.com>

* camelCase strings for enums in config files (#519)

* use camelCase for enum values in configs and update migration logic accordingly
* disable editing of CVR source rows in GUI table

* Final 1.2.0 polish (#526)

* Updates checkstyle from 8.34 to 8.36.2, and google_checks.xml.
* Creates and implements separate `checkstyle-suppressions.xml` file to suppress Checkstyle warnings that don't make sense for our project (fixes #489).
* Addresses relevant Checkstyle warnings (fixes #490).
* Addresses IntelliJ warnings.
* Pixel-pushing for Linux (get thicc, boi)
* Changes all logging to use new `fine()`, `info()`, `warning()`, and `severe()` methods.
* Gets rid of now-unused `onEditCommit` functions for the CVR Files and Candidates tables.
* Makes util classes `final`, with private constructors to prevent instantiation.
* Renames `TieBreakMode`, `tieBreakMode`, `TieBreak`, and `tieBreak` to `TiebreakMode`, `tiebreakMode`, `Tiebreak`, and `tiebreak`, respectively. This word has now lost all meaning.
* Removes unnecessary usages of `.toString()`.
* Standardizes exceptions as `exception` instead of `e` to comply with VVSG requirement on variable names (5.2.5 paragraph c) that only index loops are allowed to have single-character variable names.
* Fixes broken test.
* Minor fixes to hints.
* Get rid of the Devilish Double Spaces after periods.

* Minor cleanups (#527)

* fix README

* CDF JSON: throw if a CandidateObject can't be found during parsing.

* Cleanup CDF class (type) names as output from ResultsWriter.

* fix label

* Handle missing precinct data for older Dominion data sets #533 (#534)

* Handle missing precinct data for older Dominion data sets #533
* Add test data with missing precinct data for older Dominion data sets #533

* Fix for #536 - allow multiple CDF files (#537)

* Fix for #536 - allow multiple CDF files
Add regression test with multiple CDF files

* fix overvote delimiter logic and update test (#539)

* Fixes for bad merge between `master` and `develop`.

* Add README and LICENSE for split repo

* latest updates from develop

* update cdf output names for #584

* add dominion multi file tests

* add licence file #551

* remove test_data as submodule #612

* remove test_data submodule #612

* remove submodule reference from docs #612

Co-authored-by: HEdingfield <hylton+git@groupagree.com>
Co-authored-by: hedingfield <hylton@groupagree.com>
Co-authored-by: Louis Eisenberg <tarheel@gmail.com>
Co-authored-by: HEdingfield <HEdingfield@users.noreply.github.com>
Co-authored-by: Armin Samii <artoonie@gmail.com>
Co-authored-by: Armin Samii <armin.samii@gmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

6 participants