Talk:Client comparison chart

81.4.178.82 13:36, August 27, 2019 (UTC) louis94
 * THIS PAGE IS ORPHANED, NOT LINKED BY ANY OTHER PAGE, CLEARLY MARKED AS UNFINISHED/INCOMPLETE, and under ORIGINAL AUTHORS' RIGHTS TO FINISH according to a very OPEN methodology that INVITES contributors to partake with others. The nature of the content is incompatible with a large number of individuals single-handedly "correcting" scores without engaging in a group methodology to share and argue their observations and submit to agreeing to objective criteria over disagreed areas. Individual perceptions and interests differ greatly and the result of such would be a page constantly edit-warred every day. In order to exist, and the only way it can exist, is by submission to the most positive and fair and open process. Not doing so when invited can be counted as a recusal, self-black-balling, or self-abdication from any further input. EDIT-WARRING an orphaned incomplete unlinked page still under original authors' right to complete it, without respecting a very fair and open invitation and methodology for group collaboration and frinedly discussion, and refusing to join or participate in the community assessment/discussion process it has specified, is guilty of edit-war: please see wiki-wide policy guidance. Original authors have a right to finish an unlinked unpublished incomplete page by their own methodology without edit war. Please respect that this kind of content cannot even exist or do its USEFUL purpose if single individuals refuse to participate communally in a fair and open methodology. Choosing to not adhere to this is the same as chosing for the page to not exist at all, as surely it would have to be deleted. The result of that would damage the possibility for developers to properly assess missing features and contribute their improvements to the clients, with a net loss to the Freeciv community. Please read again that this is NOT your opportunity to market your favourite client as better and make others look worse. Accurately representing MISSING features is the chance for developers to IMPROVE your favourite client. Admitting a shortcoming gets it eventually improved. Conversely, inaccurately mis-portraying clients as worse than the Legend indicates, is an untruthful disservice to everyone. Please respect the motive and purpose of this project. You can join the offline group and be more active, or if less active, you're also invited to submit observations/suggestions/arguments/disagreements in top-down order on this page and they will be discussed fairly. As stated elsewhere, unresolved disagreements will be appended; You are WELCOME and INVITED to join the communal process to discuss and enrich the results and its accuracy, but don't assume you are God-Emperor whose opinion trumps all others while refusing discussion or exchanging information and perspectives with others. Thank you!
 * "Zero Latency Go To" was investigated and proved to be a misleading contextual comparison. Specifically, latency exists between a client and a server. All clients experienced near-zero latency when server and client were on same machine, and some degree of latency when connected to a remote server. Variability of client-server contexts should not be confused as a comparison of clients themselves. Additionally, Latency for the most frequent contexts is already subsumed under the category for Performance.
 * The visible latency is very annoying on slow networks and motivates a "-" in the Performance row. 80.244.29.176 22:43, August 27, 2019 (UTC) louis94
 * Slow networks create latency for all clients, while client/server on same machine experience virtually no latency for ALL clients. Comparing apples to oranges to specifically degrade one client relative to another is not true to the agreed esprit and methodology of this project; please read that elsewhere. This is a comparison of clients to clients, not a comparison of your preferred client A on fast local server vs your less preferred client B on remote server.
 * Should "rich go to lines" get a star by merely existing, or is there further unimplemented info they could contain if improved, that should be reserved for future star-status? Suggest upgrading current check grades to star grades unless someone can think of further improvement.
 * I have an update that I want to land in 3.1, see http://forum.freeciv.org/f/viewtopic.php?f=12&t=91203
 * This would deserve a star-grade. But showing turns deserves more than a check, so after some discussion it was agreed to graduate up to a check-plus.
 * "Overall elegance of UI" is perhaps too subjective -- is there a way to objectify into metrics such as number of clicks/presses to finish an exactly identical turn, or should this be divided into separate more objectifiable metric categories to avoid subjectivity bias? For now, this overall score is based on a weighted median of many overall scores, where the weights are admittedly somewhat subjective but are weighted in order of first to least on the following elements:  total number of clicks/keypresses to achieve actions, ease of learning, standard/universality of UI (similarity to UI behaviour in most commercial software), full ability to access anything quickly via mouse, full ability to replace mouse interaction with faster hotkeys for "power users", positioning/display/interaction/behaviour of UX/UI elements, interlinkability/navigation between remote elements (e.g., hovertext help, picking a unit inside a city and having it instantly focused on the map, etc.)
 * Should disantangle visual style, interface layout, discoverability, usability and accessibility. For example I think that the visual style of the Web client is ugly, while the layout is mostly ok 81.4.178.82 13:36, August 27, 2019 (UTC) louis94
 * RE: Desktop connecting to web.  Existence of a server with no active repo or commit history for server changes accomodating connectivity from desktop clients; as well as extended steps and methods to get clients to connect; does not qualify for star-status, but is an exciting community development. Please see Note 1 for future updating on the status of this important initiative.
 * Suggest using (3.0+) in the table since an improved UI will probably be included. 81.4.178.82 13:36, August 27, 2019 (UTC) louis94
 * This is actually a server feature, not a client feature. It should be removed from this page. Quirrelmort (talk) 17:23, August 27, 2019 (UTC)
 * Crashes/glitches - ability to finish an entire singleplayer game with less than 50% incidence of a crash-and-restart experience, is the main qualifier for a check-grade. Significantly better or worse will alter the basic check. Further modifications come from more specific glitches such as:  necessity after several hours to reload page on FCW to avoid "bogging slowness", specific but rare bugs like nuke-crash, need to do circumvention to get proper combat_rounds results (not necessary in most rulesets), etc., etc.
 * Never got a single crash in a stable version on Linux... would qualify for star-grade on this platform. 81.4.178.82 13:36, August 27, 2019 (UTC) louis94
 * Splitting by OS seems relevant to me, since users typically have whatever OS they have and don't care if it crashes on another. Will recommend later splitting this category by OS. Also helps devs know which one to fix up ;)
 * To get a star-grade for past>future city happiness status requires:  ability to instantly see without further navigation, interaction/input, whether the city is "currently" happy/celebrating/disorder as of the most recent TC AND whether its current configuration will result in that state (or a different one), were nothing to be changed at the upcoming TC; the ability in the city window to see "Celebrating" and also clearly see a majority of happy people does not fulfill criterion for this--all clients would simply get the same score and no motivation for further feature development would be indicated.  Cases such as "was-celebrating but currently not configured to continue", and all other cases must be visible, to trigger the feature functionality this is portraying, which is simply: instant recognition of all past>present>future states from a single display without needing interaction, in order to quickly act on it and manage it.
 * It sounds like all of the client display the same information then. Quirrelmort (talk) 17:23, August 27, 2019 (UTC)
 * "Map drag movement" represents the standard/universal ability in modern commercial software with modern UI/UX fucntionality, to grab a map or image and start dragging it to reposition it, with the mouse/trackpad/other non-keyboard input. To get a star-grade requires either press-and-hold-to-drag or double-tap-then-drag, or similar. Since all clients equally contain other methods (e.g., right click centering), a comparison of all clients in this area does not serve the purpose of motivating developers to implement new missing functionality from their particular clients.
 * To me a star should require edge scrolling in full-screen mode and support for touch scrolling 81.4.178.82 13:36, August 27, 2019 (UTC) louis94
 * An ADVANCED UNIT SELECTION FILTER is the ability to select units by more advanced attributes such as SENTRY, full hitpoints, full movepoints, greater offense than defense strength, and so on. No client has taken this to its full potential, to get a star-grade represents the client is superior to all other clients, and is the example for them to emulate for new feature development.
 * To get a star-grade in hotkeys requires implementation of most all common hotkeys PLUS implementation of distinct and frequently useful hotkeys not present in other clients.
 * It should require configuratility too, for people not using English keyboards 81.4.178.82 13:36, August 27, 2019 (UTC) louis94
 * Extended mouse functionality: a check or check-plus grade can come from a new user immediately observing all the standard expected user interaction functionality to be there (map-drag, shift-click selection toggling, etc.). A star-grade means (most) all of this is there and the client is the current best example for other clients to emulate--but does not mean it can't be extended even further.
 * Alternate cursor/numpad input does not mean MAIN cursor/numpad movement. It means an alternate exists for the substantial minority of users who do NOT have numpads. Mapping 2D cardinal directions onto a 1D strip of numbers from 1-9 at the top of the keyboard is indeed an alternate to using a numpad. Under the Legend it qualifies exactly as minimum or poor, thus, an X- grade. Keys such as alt + letter keys arranged as a virtual numpad, qualifies for a star-grade.
 * In some keyboards (mine) you need Shift to access the number keys, and their meaning is only intuitive if you have the keypad in mind. Would suggest using a (configurable) WASD-style layout instead. 81.4.178.82 13:36, August 27, 2019 (UTC) louis94
 * Map Zoom Level: to the user who gave a star-grade to GTK for map zoom. Please note the manual gives no indication of any controls for map zooming; and the developer response to this question: http://forum.freeciv.org/f/viewtopic.php?t=66166   A star-grade would mean absolute minimal keys/clicking to instantly and dynamically alter the zoom level on the fly. If removal of the star-grade was an error, please write so here and give the exact hotkey/click(s) to achieve it. Thanks!
 * Star-grade should be awarded only if the standard shortcuts are implemented: Ctrl+mouse scroll, Ctrl++, Ctrl+-, pinch. 81.4.178.82 13:36, August 27, 2019 (UTC) louis94
 * Orders buttons: the mere presence of orders buttons does not qualify for star-grade, only a check-grade.  To gain higher levels, consider:   Iconography: ease for the eye to immediately discriminate the graphic from other graphics. Completeness: buttons for all legal orders. Contrast/ordering: colours/sequence/grouping/etc., make mental processing much quicker for finding the button wanted. Extended functionality: ability to customize which buttons show or don't show, collapse/remove buttons panels, extend/collapse between all/frequently used, etc. A star grade would require best-in-class functionality for all these OR close enough in all these while implementing extra functionality of great worth and frequent use, that other clients lack. IF the orders buttons are: missing some legal orders, have poorly contrasting iconography, are all the same colour and/or lack grouping/contrasting, then please observe the Legend grading system and respect this is earning of a grade somewhat reduced from a standard check-grade, not a star-grade.
 * These criteria are designed to give the Web client a star-grade. 81.4.178.82 13:36, August 27, 2019 (UTC) louis94
 * To the user who added star-grades for mobile apps for SDL and QT: thanks(!) for pointing out the availability of these on mobile devices. We believe that the functionality of these whether via web browser or as standalone mobile app, all have considerable room for improvement. All three were downgraded to check-grade in the interests of accuracy and perhaps motivating further development and improvement for all three.
 * August 2019 current stable and dev GTK version theoretically works with combat_rounds but has display error of not updating unit stats post-combat. Save/reload or Sentry/unsentry are known workarounds.
 * This would actually be a server issue, not sending updated unit packet. Quirrelmort (talk) 20:16, August 27, 2019 (UTC)
 * Incorrect. Packet gets sent. Fix was implemented in the client.
 * Confirmed that native clients display combat results correctly. Quirrelmort (talk) 23:30, August 27, 2019 (UTC)
 * What is a "Server GO TO" bug? Quirrelmort (talk) 17:23, August 27, 2019 (UTC)
 * It is a bug where a GO TO is falsely reported by the server to the client as illegal (thus client logic "faithfully" disallows it), yet when using manual cursor keys it is 100% legal and the server will indeed perform the manual steps to enact it.
 * That is purely an error in the server-side pathfinding that only the webclient relies on. Quirrelmort (talk) 22:42, August 27, 2019 (UTC)

Objectivity
Some entries in the table cannot be made objective. Namely:

80.244.29.176 22:08, August 27, 2019 (UTC) louis94
 * Overall elegance/simplicity of UI: primarily opinion-based
 * Please see notes on giving how semi-objective status can be given to some of these. Ultimately, small minority dissenting opinions can be footnoted and appended and given a chance to present objective and subjective arguments. 80.244.29.176 22:08, August 27, 2019 (UTC) Zarah Witchcraft
 * Extended/improved graphics included for most popular tileset: we don't have tileset statistics; "improved" is primarily opinion-based
 * amplio2-derived is most commonly used to such an overwhelmingly large degree, that debate to remove this info has to be questioned as uncollaborative. The chart becomes more informative to all demographics, not when amplio2-haters remove this info, but by including info about a specific client that has extra features for non-amplio2 tilesets. 80.244.29.176 22:08, August 27, 2019 (UTC) Zarah Witchcraft
 * Extended/improved sound effects included: "improved" is primarily opinion-based
 * if, among those who play with sound ON, the great majority says "wow this is so much better", then we can take this type of argument to be coming more from a client partisan. Please note the principles and constitutional spirit of the chart is not to argue or fight to present one client as better or not in its current state. The purpose is to actually motivate client improvements. Hiding the fact that an improvement exists that a supermajority of people like, in order to make a client without that feature seem to compare better -- this is anathema and completely counter to the purpose of these charts. The proper purpose would be to get a developer of the other client to go acquire the open source content and include it in that client, to bring up its quality. 80.244.29.176 22:08, August 27, 2019 (UTC) Zarah Witchcraft
 * Overall scrollable/readable: primarily opinion-based
 * Scrollability and readability are primarily objective.  Presence of scrolling is a fact/non-fact, as is whether it is smooth and reliable or jump/glitchy or requires larger number of clicks/presses to interact with.  Many of the same arguments apply to readability, which is number of clicks/presses to get to an info panel, number of content items displayed and compared in the panel, and professionally accepted dogma on contrast/layout and other Gestalt factors known to perceptual sciences to increase/decrease the ease of gathering and processing displayed information. 80.244.29.176 22:08, August 27, 2019 (UTC) Zarah Witchcraft
 * Number of clicks or presses for frequently performed tasks: "frequently performed tasks" is vague and depends on the player and type of game
 * If the argument is that client developers and client users should have this information censored from them because it's vague, that is counter to the esprit of the project. Some degree of vagueness and good faith fair treatment of it is a requirement for this kind of project. Everyone can agree that selecting and moving units, investigating what's inside a city, setting a research target, responding to a message or diplomatic offer, are frequently performed tasks. 80.244.29.176 22:08, August 27, 2019 (UTC) Zarah Witchcraft
 * Criteria based on the number of clicks don't apply to all types of entries, because some less common actions will always require several clicks. This requires clarification.
 * For loose descriptions of client as a whole, it can be thought of loosely as the "median experience" for "median action performed". In other cases where it is referring to a more specific action or interaction, it can be specified exactly. 80.244.29.176 22:08, August 27, 2019 (UTC) Zarah Witchcraft
 * The following criteria are primarily opinion-based: Superior, inferior, difficult to see or use well. They need to be replaced.
 * They "need" to be replaced if you and others who join the project want to turn this into a two year 4000 man-hour academic peer-reviewed treatise. Otherwise, it can generally be accepted that 6 clicks and 3 seconds to perform X is inferior to 1 click and 1 second to perform X. The intent is NOT to hide from users that 6 and 3 are inferior to 1 and 1 so that they are misled into believing a certain client is not inferior when performing X. Rather, the purpose is to motivate development on all clients by allowing a quick survey of those who favour certain clients to easily see where their favourite client is inferior and can be improved. 80.244.29.176 22:08, August 27, 2019 (UTC) Zarah Witchcraft
 * If objectivity is a goal, make all entries Yes/No like in a typical feature comparison chart https://en.wikipedia.org/wiki/Comparison_of_email_clients 80.244.29.176 22:08, August 27, 2019 (UTC) Quirrelmort (talk) 14:07, August 27, 2019 (UTC)
 * Objectivity has two meanings. The second and often more important meaning is Objective Reality. Reducing objective reality to yes/no censors important relevant details: in this sense, yes/no is anti-objective or counter-objective.  If two clients both have Feature X, but one of the clients requires navigating 4 levels down into a Preferences menu that many users will never find and most users will only do very rarely, whereas another client can do it with a single hot-key, then it is decidedly unobjective to show a simple YES/YES comparison. This begs a question of political/partisan censoring or skewing of content as if the purpose is to market clients as products to consumers and the "manufacturer" of a certain client wants to hide areas that are not as good. Please carefully read the purpose of this project, which is to factually portray infos to developers so that they can easily assess areas to improve their clients, and to factually portray these infos to users who can decide based on their own individual tastes and preferences. 80.244.29.176 Zarah Witchcraft

Additional reply

 * Pure epistemic objectivity, as if peer reviewed Physics, is NOT the raison d'être of the client comparison. Please read what it is.


 * Maximal objectivity when possible is preferred and suggestions where to improve it are welcome. In cases where it isn't achieved and subjectivity creates dissent among users who WANT to show flaws of their favourite clients in order to MOTIVATE improvements to those clients, a mechanism is built-in for splitting such information. Finally, in cases where someone simply wants to impose minority subjective opinion to monopolise their individual ppinion that their client as-is, is superior to others; they are incompatible with the goals of this page and will have use some alternative space to voice themselves.

80.244.29.176 Zarah Witchcraft
 * Let's remember that there can be the case of a father claiming his ugly daughter is more beautiful than an instagram model with a million followers, and he claims beauty is subjective; nevertheless one million to one becomes an objective metric. Sciences have evolved many mechanisms for "objectified subjectivity" to transcend secondary school granularity for filtering by "yes/no" and "fact/opinion".

Suggested rewording of the legend
I suggest to use the following rules to rate entries: 80.244.29.176 23:03, August 27, 2019 (UTC) louis94