Talk:BSicon/Renaming/ABZ

Diagonal junctions

1 branch

1-branch 45°
Old	New
(5006)	(ABZ1+3r)
(5002)	(ABZ1+3f)
(ÜWl+r)	(ABZ2+4g)
(5004)	(ABZ2+4r)
(ÜWr+l)	(ABZ3+1g)
(5005)	(ABZ3+1l)
(5003)	(ABZ4+2l)
(5007)	(ABZ4+2f)

1-branch 90°
Old	New¹	New²
(ÜWgo+l)	(ABZ+1)	(ABZg+1)
(ÜWgu+l)	(ABZ+1u)	(ABZg+1u)
(ÜWgol)	(ABZ2)	(ABZg2)
(ÜWgul)	(ABZ2u)	(ABZg2u)
(ÜWgor)	(ABZ3)	(ABZg3)
(ÜWgur)	(ABZ3u)	(ABZg3u)
(ÜWgo+r)	(ABZ+4)	(ABZg+4)
(ÜWgu+r)	(ABZ+4u)	(ABZg+4u)
(¹) As proposed by Useddenim on 2012.05.18 and further discussed. (²) As modified by Axpde on 2013.02.16 09:33.

You omitted the "g" to indicate the straight track as in  (ABZgr+l)! a×p_de^Hello! 09:33, 16 February 2013 (UTC)

Who is «you»? And did took you from May to February to notice these renaming suggestions? In the thread below (Talk:BSicon/Renaming#Still about ÜW filenames) seems that nobody agrees that this "g" is necessary. I, for one, find it confusing and misleading: If it is necessary, how does an ABZ look like without it? What does "g" add to these icons? -- Tuválkin ✉ 11:54, 16 February 2013 (UTC)

Still about ÜW filenames

I noticed this — i.e. that somehow   (uABZ+4) was once defered to, and remade into a reditect to,  (uABZg+4). What’s that "g" and why do we need it? Gegen to, what?

I understand that many 45° still have their ÜW-names and it is hard and slow to change them to 1234-names because many have really vast usage. But seems that renaming   (uABZ+4) to an unstandard name is a step backwards, or even backwards & sideways.

Opinions?

-- Tuválkin ✉ 15:55, 7 January 2013 (UTC)

"g" to indicate "straigth track". a×p_de^Hello! 20:30, 27 February 2013 (UTC)

You'll have to ask Axpde (talk · contribs) why he made the move; but it appears that he's been away from BSicons for too long, as I thought the question of g/no g had been settled long ago when the topic of renaming the ÜWABZ icons was originally discussed. Useddenim (talk) 18:05, 7 January 2013 (UTC)

Settled sounds right. I’ll boldly (and slowly) go at fixing these, then. -- Tuválkin ✉ 18:54, 7 January 2013 (UTC)

 (uABZg+4)→  (uABZ+4)  Done -- Tuválkin ✉ 20:19, 15 January 2013 (UTC)

Look no further than Diagonal junctions above. Useddenim (talk) 04:10, 8 January 2013 (UTC)

Ermegerd, I just finished manually fixing the redirects   (ÜWgol)→  (ABZ2): It was not easy, they were more than 150. Now, your turn to tackle another like it, guys. -- Tuválkin ✉ 20:52, 9 January 2013 (UTC)

Of course Axpde had to come over and vandalize my work. He doesn’t have much free time or patience any more to do any actual work, but to keep pulling his string in order make sure all his ideas are left unchallenged — well, for that he always finds himself a minute. Sad. -- Tuválkin ✉ 11:41, 15 February 2013 (UTC)

For several years I'm working now on BSicons to have them consistently named. I created thousands of them and edited tenthousands of pages manually to get rid of inconsistently named duplicates!

And now you create icons named according to your personal point of view before introducing your ideas to the community. And you call me a vandal and a troll ?!? You should calm down immediately! a×p_de^Hello! 12:14, 15 February 2013 (UTC)

Paraphrasing Churchill, tomorrow I’ll be calm. Your work concerning BSicons doesn’t give you any edge over the dozen other editors similarly envolved in this matter, me included. Your claimed concern about stability is shared by all of us, and would be trivial to find instances of me and others working hard to keep naming consistent. What sets you apart is the hamfisted approach you take, using your admin abilities to lock files and thus salting the ground for discussion and experimentation, and at the same time you yourself coming to discussion with an aloof, tone-deaf, and unconstructive approach. That doesn’t sit well, especially when you are insisting on ilogical notions — such as in the GRENZE matter, or in your bizarre approach to categorizing (against what is done in Commons, really). So, when I acuse you as above, that’s not an insult, but a statement of fact (as I can see it), just like is not an insult to say that bats suck, or to call a puppy a son of a b¡tch. As for your accusation that I introcuce new naming practices without discussing them, done under a byline that is exactly an invitation to discussion («ongoing experiment») — now that shows that you need to reevaluate your behaviour, as much or more than I need rein in my bad temper. -- Tuválkin ✉ 04:50, 16 February 2013 (UTC)

Ok, just to see it clear:

1. You say it's not ok to rename an icon to a consistent name where we had an agreement about (i.e. terminal stations).
2. But you say it's ok to post an idea an then proceed although noone explicitely agreed with your ideas?

a×p_de^Hello! 09:33, 16 February 2013 (UTC)

«Just to be clear», this:

1. I don’t agree with that agreement (terminal stations being "l"/"r" instead of "a"/"e" (±"q"), as a logical outcome of the whole "q" affair) and explicitly chosen a seldom used icon color to present an experiment and illustrate my point. You arrive from the mists and played your usual jackbooted shtick, as you did in the GRENZE affair.
2. What are you talking about? If it is the 45° ABZs, you are the one whose viewes nobody else agrees with — that "g" seems to be unnecessary, incomprehensible, confusing, and misleading. And I only started renaming these months after Useddenim said go-ahead.

-- Tuválkin ✉ 12:28, 16 February 2013 (UTC)

Months? I see that Axpde has just changed   (ABZ2) (525 uses) to  (ABZg2) (84 uses) one year (356 days, to be precise) after I changed the name from   (ÜWgol), without any problem until now. Frankly,

a) the superfluous g is unnecessary: ABZ = junction 2 = to corner 2;

b) these changes are being made arbitrarily, without discussion and consensus; and

c) if Axpde wants to do something useful, he should correct  (ABZlg) and  (ABZlg).

Useddenim (talk) 16:39, 16 February 2013 (UTC)

The months I meant are from May when you said go ahead about renaming the ÜWg(o/u) icons to ABZ+number, to January when I renamed  (uABZg+4) to   (uABZ+4), as discussed here. -- Tuválkin ✉ 04:07, 17 February 2013 (UTC)

Meanwhile, this: Axpde says that «the "g" to indicate "straight track" was missing». So "g" indicates a straight track now, not (only?) gegen. (Ain’t it cool? I mean, really!) In the last couple hours Axpde renamed those icons I was working on after Useddenim’s suggestion — he renamed the icons, but didn’t bother with the global usage: The icon   (ABZ2) (now renamed  (ABZg2)) alone has more than 250 750(!) uses in several projects — all these pages are now showing damaged diagrams. All because Axpde woke up after eight months and decided to rename icons against what we had agreed. -- Tuválkin ✉ 12:11, 16 February 2013 (UTC)

Newsflash: Axpde thinks that these ABZs that go under the corner show not bridges, but tunnels — see example. Now, that’s one fine case of previous discussion with the other envolved editors, right? And also note how he has been frenetically changing all names in one go, not bothering with a few thousand pages damaged across a dozen wikipedias, unlike what we usually do — renmaing and fixing use for each icon, slowly for minimum disrupt. This is frankly disstressing. -- Tuválkin ✉ 13:08, 16 February 2013 (UTC)

It is a «small change», he says… But I say it is a big mistake, because those are not tunnels: Just cp. the shapes of (bridge) and (tunnel). -- Tuválkin ✉ 13:22, 16 February 2013 (UTC)

Yes, you are right. the "u" -> "t" change was no good idea :( I'll change it back ... a×p_de^Hello! 20:37, 27 February 2013 (UTC)

Is it you that delete ABZ2,ÜWgor,ÜWgo+l and ÜWgo+r in Chinese?

Moved from en:User talk:Axpde.

I am a Chinese contributer.Today I find ABZ2,ÜWgor,ÜWgo+l,ÜWgo+r are deleted,and many Chinese pictures,such as zh:User:江漢思歸客/Sandbox,cannot be read normally.Is it you that do so?What should I and other Chinese contributers use from now on?I wonder if you are willing to help me fix them by robots.If so,how soon?And if not,does it mean we have to fix them by ourselves?

Sorry for my poor English.Thank you.

江漢思歸客 (talk) 16:13, 16 February 2013 (UTC)

I have to strongly oppose your Axpde's renaming of the ABZ# icons without any discussion, especially as they began to be introduced over a year ago. Furthermore, naming was also discussed in August 2011 (Talk:BSicon/Renaming/Archive 3#Names stuff), so this should have been considered "settled". Useddenim (talk) 21:50, 16 February 2013 (UTC)

Sorry I have only edited Chinese wiki till now, and as far as I know,I can only use thse 4 icons before yesterday.I know nothing about so-called "renaming".Now they don't exist any longer.The problem at present is my work can't be read,and what we should use in future.

江漢思歸客 (talk) 02:19, 17 February 2013 (UTC)

Dear 江漢思歸客, Useddenim is addressing Axpde, not you, in the matter of this problem, as it was he (Axpde), who caused the problem in the Chinese wikipedia you complain about, as well as the same problem in several other wikipedias — as discussed above. The renaming done by Axpde is bad in 3 ways:

• was not discussed previously, and indeed, goes against previous discussion;
• was done in one go, affecting eight icons and thousands of uses, and was left untended, instead the usual gradual single icon change followed by swift fixing of all its uses;
• it is obviously wrong.

That’s what I also strongly oppose. -- Tuválkin ✉ 03:50, 17 February 2013 (UTC)

Axpde (talk)'s comments, verbatum, from 7 June 2011 (with respect to Renaming "legende" on this very page!):

"concensus" does not mean you or me, even commons is the wrong place, we should at least ask the two biggest projects w:de: and w:en:!

(My emphasis added.) Useddenim (talk) 12:06, 20 February 2013 (UTC)

This matter needs to be not forgotten. Just today there was this:

(Move log); 12:48 . . Axpde (talk | contribs) moved page File:BSicon eABZ2.svg to File:BSicon eABZg2.svg(consistent naming scheme)

There is nothing «consistent» in adding a "g" for which only Axpde sees the need. This renaming goes against consensus among the users’ community, and even if there were a good reason for it, Axpde is failing to allow it to be discussed. There were previous instances of this behaviour, but never in such a massive scale: Never so many icons, never so widely used, never against such a clear opposition from others. This is admin abuse and shall be reported. -- Tuválkin ✉ 14:27, 13 March 2013 (UTC)

Hear, hear!! (But who will listen?) Useddenim (talk) 00:42, 14 March 2013 (UTC)

Things are getting out of control…!

Do we really need five different ways of naming what is essentially the same geometry:
 (ABZg+4)   (xABZ+4)   (eÜWgo+r)   (ABZg+4t)   (xÜWgu+r) ?!!!
Useddenim (talk) 00:59, 6 March 2013 (UTC)

Well, you can break that up thusly:

•  (ABZg+4) vs.   (xABZ+4): Hijacked by Axpde 2 weeks ago, against many months of usage rooted on solid dicussion. He may go on adding that useless confusing "g" to all such icons.
•   (eÜWgo+r): Old name of the above still around, showing that there’s a lot of renaming work to be done still.
•   (ABZg+4t) (this is a different icon type, with a bridge, not flat): You forgot   (ABZ+4u), which was the agreed name for this icon, before the mentioned hijacking took place. Later on, Axpde changed his mind and doent’t think these should named as tunnels, after all…
•   (xÜWgu+r): This is the old name of the above, ditto.

:-\

-- Tuválkin ✉ 22:20, 6 March 2013 (UTC)

Ok, I just reverted my move 'u' -> 't'. Now I'll try to tidy things up ... a×p_de^Hello! 12:04, 14 March 2013 (UTC)

1. ABZg+1t, ABZg2t, ABZg3t, ABZg+4t are completely exchanged and deleted.
2. xÜWgu+r was moved to  (xABZg+4xu), completely exchanged and deleted.
3. all e/x/ex ÜWg-icons are completely exchanged and deleted.
4. ÜWgor, ÜWgo+r, ÜWgo+l and ABZ2 are exchanged except for some occurances in :jp:wp (about 40~50 pages left).

Will be done soon! a×p_de^Hello! 13:21, 16 March 2013 (UTC)

2 branches

2-branch 90° Existing Proposed   (5021)  (ABZq+13)   (5022)  (ABZq+24)   (5023)  (ABZ+13)   (5024)  (ABZ+24)

2-branch 45° Existing Proposed   (5011)  (ABZ3+1gf)   (5012)  (ABZ3+1lr)   (5013)  (ABZ2+4gf)   (5014)  (ABZ2+4lr)

Anyone have any thoughts? Useddenim (talk) 12:47, 18 May 2012 (UTC)

I added suggestions for the missing ones, «2-branch 45°». Everything else seems okay. -- Tuválkin ✉ 23:13, 18 May 2012 (UTC)

Are diagonal junction IDs sorted up? Thank you for all. I am writing for the uploader whose home is JAWP, once bringing a mess here. We in JAWP use IDs 5xxxs for test icons. I do not mean it is exclusive. --Maxima m (talk) 11:29, 5 June 2012 (UTC)

 Done except for 5022/ABZq+24 (wikimedia ate the icon, and I have no idea how to restore it...) Useddenim (talk) 10:43, 13 June 2012 (UTC) Useddenim (talk) 22:50, 13 June 2012 (UTC)

l & r order

split from Talk:BSicon/Renaming#kKRZ: Naming conflict

P.S.: Conc. "lr" vs. "rl" - there is no "rule", but to have it uniformly I'd prefer alphabetical order, i.e. "lr". a×p_de^Hello! 13:02, 24 February 2013 (UTC)

Choosing "lr" instead of "rl" seems to be the usual practice — it should be spelled out as a rule, or else ppl will assume a given icon doesn’t exist when it doesn’t up show up if called by the equivalent name. Also necessary to fix the tables at en:Wikipedia:Route diagram template/Catalog of pictograms/compound junctions and rename any stray icons. -- Tuválkin ✉ 21:04, 25 February 2013 (UTC)

If we are using alphabetical order, then even more so that corner numbers should be in sequence. See this for Axpde's latest aberration. (If both 1 and 4 were ex, then it would be  (exABZ+14), not   (ABZ+x14). See also  (ABZ+1x4).) Useddenim (talk) 16:32, 6 April 2013 (UTC)

"aberration" ... as kind as usual! Think nothing of it. Besides, Tuvalkin's used harsher language. 8-) Useddenim (talk) 21:44, 6 April 2013 (UTC)

As I already said above the order is: "coming-from" x "coming-from" + "running to" x "running to". See e.g.  (ABZrxl),  (ABZ+rxl) and others! a×p_de^Hello! 20:14, 6 April 2013 (UTC)

So you've just decided that "x" (the 23rd letter of the alphabet) supersedes other alphabetization. Counter-intuitive to me. l/r, 1/4, etc. are the geometry which doesn't change while x is merely a state modifier: all the more reason the order should remain unchanged while x moves where necessary. Useddenim (talk) 21:44, 6 April 2013 (UTC)

The prefix order is u-e-x-m-f/g-b/c/d-v-h/t which isn't alphabetically as well! It was never said that suffixes has to be in alphabetical order! So "x" doesn't supersede others, it negates a certain direction! a×p_de^Hello! 16:00, 7 April 2013 (UTC)

Actually I agree that "x", which always goes before all the suffixed parts that are ex-ed. We need to be able to separate the 3 situations that happen when the branches of  (ABZglr) become disused, leading to -xlr (all disused)-lxr (right one disused) and -rxl (left one disused). Circeus (talk) 05:20, 7 April 2013 (UTC)

Consider

All in use	1 out of use	2 out of use	All out of use
(ABZglr)	(ABZglxr)  (ABZgrxl)  (xABZglr)	(ABZgxlr) ?  (eABZglr)	(exABZglr)

Useddenim (talk) 12:29, 7 April 2013 (UTC)

The "e" prefix stands for "main track in use, facility off use", the "x" prefix "main track use, facility in use". That's why it's  (eABZglr) instead of ABZgxlr. a×p_de^Hello! 16:00, 7 April 2013 (UTC)

Right. e- slipped my mind (thought I'm sure there are kooky cases somewhere, like a {BS-q|kKRZorl+rl}} with partial closures, that would require xlr or an equivalent). The point is, I think -lxr/rxl are better or at least less liable to misinterpretation than -xlr/-lxr, but I'll admit it's a personal choice. There are arguments both ways. My impression was that the current accepted system is the middle x with reversal where required and I'm not sure changing the system is really needed. With that caveat, I think that when in presence of all of -f/-g/-l/-r, the order should be -fg then -lr (-f and -g always first, in alphabetical order, then -l/-r also in alphabetical). Circeus (talk) 17:28, 7 April 2013 (UTC)

Good summary. Can this be the last word on the subject? Useddenim (talk) 00:24, 8 April 2013 (UTC)

Back in 2006 when Bernina created  (ABZlf),  (ABZlg) and so on the BS-templates were meant to show a single line from start at top to end at bottom. At this time adding an "f" for "driving direction" and "g" for "opposite direction" was the most simple idea, but it wasn't added in alphabetical order – and nobody intended to enhence the set of icons to what we have today. But soon it was clear that more complicated icons were needed that's why I dropped "f" and "g" and introduced the "+" to represent tracks "coming from ..." a×p_de^Hello! 19:21, 7 April 2013 (UTC)

P.S.: Bernina's original design looks pretty much like this  (ABZlf)! a×p_de^Hello! 19:29, 7 April 2013 (UTC)

Former ABZ3

moved from User talk:Axpde#Former ABZ3

Hi, you seem to be preparing migration from "ABZ3"s to "ABZq"'s; however, as far as I understand, the titles you chose are completely wrong. (The previously existent  (exABZql)'s should likewise be swapped in places.) Isn't it the Circeus's proposal that should be followed? Me and Tuvalkin has recently moved a couple of ABZ3's to new titles, and we followed that proposal. Cf. this table. YLSS (talk) 09:23, 26 March 2013 (UTC)

l	r	+l	+r
ABZql	ABZqr	ABZq+l	ABZq+r
ABZqlxr	ABZqrxl	ABZq+lxr	ABZq+rxl
ABZql+xr	ABZqr+xl	ABZqxr+l	ABZqxl+r
ABZql+xl	ABZqr+xr	ABZqxl+l	ABZqxr+r
STRl+xr	STRr+xl	STRxr+l	STRxl+r
ABZglxr	ABZgrxl	ABZg+lxr	ABZg+rxl
ABZgl+xr	ABZgr+xl	ABZgxr+l	ABZgxl+r
ABZgl+xl	ABZgr+xr	ABZgxl+l	ABZgxr+r

First of all thanks for indirectly agreeing with the fact that "ABZ3" needs to be renamed.

The names I choosed are 100% consistent with all other junctions I uploaded so far (see right):

The logic of this naming scheme is:

• l means track running to the left (as seen in direction of travel, top to bottom)
• r means track running to the right (as seen in direction of travel, top to bottom)
• +l means track running from the left (as seen in direction of travel, top to bottom)
• +r means track running from the right (as seen in direction of travel, top to bottom)
• g means track running straight (German: "geradeaus")
• q means track running aqross (German: "quer")

Concerning User:Circeus/BSicon renaming/Simple junctions#2-way branchings I have to admit that I don't understand the logic behind his naming proposal. At least it is not as easy to understand as what I wrote above ... a×p_de^Hello! 10:00, 26 March 2013 (UTC)

So, 2 yrs after me and Circeus using (slowly and carefully) "q" for this kind of icons under a rationale that results in totally opposite names, you arrive and, without even discussing it, start renaming and causing the chaos we’re seeing. Really, are you trying hard to lose your admin privileges, is that it? -- Tuválkin ✉ 10:39, 26 March 2013 (UTC)

And we still don't have complete consensus whether q should be used for any track running across, or only if it has 180° rotational symmetry. Useddenim (talk) 11:11, 26 March 2013 (UTC)

Yes, that’s why me and Circeus didn’t go ahead renaming ABZ3 (which nobody thinks is a good name, btw), only adding (me) here and there a few experiments, such as the other-color icons YLSS mentions. -- Tuválkin ✉ 11:28, 26 March 2013 (UTC)

(with ec)

Pardon, q was always the suffix to indicate tracks running across, the first icons I uploaded with the q-suffix date back to 2007! I do not know your interpretation of the q-suffix, but it is nowhere documentated.

The q=across is documentated on the following pages:

1. de:Wikipedia:Formatvorlage_Bahnstrecke#Bilderkatalog
2. Commons:Icons for railway descriptions/Policies
3. es:Wikipedia:Plantillas de esquemas de líneas#Sufijos
4. ru:Википедия:Маршрутные шаблоны#Суффикс
5. zh:Wikipedia:铁路系统标示/图标列表#後綴

a×p_de^Hello! 11:39, 26 March 2013 (UTC)

Axpde, this is not about any alternative proposal, nor even about yours in itself. It is about how you set off to do things about which there is no consensus about and even sometimes against the consensus. The documentation you list (of which only wp:de and wp:en are original and reflect a consensus, the rest is just half assed translations, now mostly abandoned and obsolete) is old news and as you know very well what has been parroted ad nauseam is that "q" should be used only for icons with rotational symmetry at 90°, which is not the case of ABZs. This rationale you present now goes against what Useddenim and you (and others) ever said about "q": Change your opinions if you want to, but don’t try to say this follows from established practice. -- Tuválkin ✉ 12:22, 26 March 2013 (UTC)

Tuválkin, maybe you are too new to remember, so I gonna tell you about "q"-suffix. Back in 2007 I introduced it on :de:wp: to indicate tracks running across while someone else introduced an "H"-prefix on :en:wp: for the same use. Obviously it arose a long and widely participated discussion about which one suits better and finally the "q"-suffix was regarded more logical (esp. when changing the "-ELEV"-suffix to the "h"-prefix). And since this day the "q"-suffix indicates a track running across. Just it, nothing less and nothing more. Esp. no implicit order to turn an icon by 90° in whatever direction to get the result!

To name it exactly: Prefixes always describe the state (u,e,x,m,f), the form (b,c,d,v) or the kind (h,t) of an icon, suffixes always describe the direction (a,e,g,q,l,r,1,2,3,4) or the level (o,u) of the track!

So if you and Useddenim decided to find another meaning for the "q"-suffix, it's not my problem. I created hundreds of BSicons following this old consense that has been written down on several pages on different projects. And to parrot myself your use of the "q"-suffix has never been documentated anywhere! a×p_de^Hello! 11:02, 28 March 2013 (UTC)

You’re so full of yourself you don’t even understand that I and Useddenim have antagonic different ideas about how to use "q". As for «documentation», let me repeat that what you so pompously refer to as such are mostly obsolete instructions that has been superceded by later decisions found in discussion such as this. My idea of "q" (not really only mine, as it is shared with Circeus, for one) is not «documented» because it is a proposal — it should be widely accepted before it is added to those instructions pages (if they ever get updated): But of course a suggestion for discussion is something you cannot wrap your head around. You only agree to explain the reasons behind your actions when everybody is is really pissed off, as it is the case now. Actually Axpde, you can harp about how much me or others are «new» to BSicons and how you clobbered out the very first  (STR) with a silex hatchet, but it is easy to gauge how much work each of us has put in the project and frankly there’s only one thing for which your stand from the BSicon crowd… -- Tuválkin ✉ 15:05, 2 April 2013 (UTC)

And for what it's worth, my definition of "q" usage is documented (here, for example) because a) it does not conflict with your original definition, and b) is a subset of Tuvalkin/Circeus’ proposal. Useddenim (talk) 21:41, 2 April 2013 (UTC)

The logic in Circeus's proposal is like that (at least that's how I understood it, and it was more straightforward to me than other versions):

• q means track running across, and the whole perception of directions in this case goes as if we follow this track left-to-right, that is facing right. This is why "q" stands closer to the root than "l", "+r" etc. - first we rotate our axes and then we add branches.
• l means track branching to the left (as seen in direction of travel, left to right)
• +l means track joining from the left (as seen in direction of travel, left to right)
• r means track branching to the right (as seen in direction of travel, left to right)
• +r means track joining from the right (as seen in direction of travel, left to right)
• I won't argue about g (since personally I agree that it should be present, but that should be (or was) decided elsewhere).

-- YLSS (talk) 11:58, 26 March 2013 (UTC)

(with 2x ec)

YLSS, let me cite the most important part of your answer: "at least that's how I understood it"

Your explantion for the q-suffix is far more complicated as all other rules conc. BSicons! Most important, why left-to-right and not right-to-left? We just managed to have all icons seen in direction of travel, i.e. from top to bottom of the icon, and now they're proposing a complete new meaning? O yes, it wasn't me who proposed a new meaning, and this was nowhere officially introduced! That's why I will continue with the old and 100% consistent naming scheme! a×p_de^Hello! 12:34, 26 March 2013 (UTC)

Well, I risk parroting, but I didn't really invent or propose it. Such an explanation coalesced inside my head from the whole lotsa talks I've read here, like that "q" means rotating 90 degrees etc. And since at least three persons agreed more or less with it (that is, Circeus, Tuválkin & me), I guess it is reasonable. Although yes, it isn't simple, so no bad feelings if we agree on something else. YLSS (talk) 12:51, 26 March 2013 (UTC)

The problem with "g" for "geradeaus" is that it is also the "g" for "gegen" and, while it may work for ABZs, it will come crashing down unexpectedly when applied to other kind of icons. We seen this kind of rickety cardhouse nomenclature fall apart before. -- Tuválkin ✉ 12:22, 26 March 2013 (UTC)

(with ec)

There is no problem, it's only used with "ABZ" or "STR" in combination "lg" and "rg" or with "STRg" to incicate the opposite direction of travel. a×p_de^Hello! 12:34, 26 March 2013 (UTC)

(ec) That said, I do see the logic in your scheme as well, and TBH theoretically it would be superior - but only if we implement it at a far larger scale. It could go like that: we have several building blocks:

 (g),  (q),  (l),  (r),  (+l),  (+r),  (2),  (xg),  (+xr),  (+x3), etc.

and we compound them together to get what we need; only I would've used solely "STR" root then. Thus we would have:

 (STRg),  (STRgr),  (STRgl+xr) etc.

That would be really simple and algorithmic (a vital point indeed that we nearly lack). However, I'm not yet clear on whether all "+"'s should (could) be grouped together, as well as "x"'s. (Also, in this case I would've switched the semantics of "l" and "r" to mean left and right side of the screen, so that we completely get rid of "directions" and practise WYSIWYG.) YLSS (talk) 12:25, 26 March 2013 (UTC)

Some points:

1. YLSS's characterization of my scheme is entirely accurate. It basically means icons are rotated 90° counterclockwise by adding -q. That is  (ABZql) becomes  (ABZq+l) because it's a rotated  (ABZ+l) (that the -lg/rg/rf/lf suffixes are still around with not even redirects is another kettle of fish, though)
2. IIRC (it goes way back now), me and Tuvalkin Useddenim^{128.205.48.118 14:04, 27 March 2013 (UTC)} mostly introduced -g (and -f) as a way to deal with the three way icons by introducing a WYE root, currently used mostly for a couple diagonal icons like  (WYE1+rf).
3. Whatever happens of -g for "through line", it—or it's replacement, since a way is needed to tell apart things  (ABZgl+lr) and  (ABZl+lr) (my proposal was _ for icons with the missing through line)—it should be used as sparingly as possible.
4. Although I do like the idea of a "modular" system, I think the current one does a relative good job of that by basically "presuming" the -g icons (i.e. the through line is stated in the root, not the suffix). Although the system does have the advantage of "solving" the _LEGENDE issue along the line by turning  (BHF legende) and  (BHF) into  (BHF) and  (BHFg), I suspect such a reworking of the system might be beyond our hands at this point (not to mention the issues of likely opposition on the projects).

Circeus (talk) 06:18, 27 March 2013 (UTC)

 (ABZ23) vs  (ABZg23): The only situation where g is needed in an ABZ.
But then consider:
 (ABZg+14),  (WYEf+14) and   (WYEg+23).
I think this perfectly illustrates the (unnecessary) confusion of over-using the "g" suffix. Useddenim (talk) 18:07, 27 March 2013 (UTC)

First of all: What's the sense of the "+" in   (WYEg+23)? See  (ABZ23) - without "+"!

Furthermore: If  (ABZ23) is regarded ok (I do think so), why does   (WYEg+23) need a "g"? IMHO it should be   (WYE23)! a×p_de^Hello! 22:05, 27 March 2013 (UTC)

P.S.: We have ABZ for "normal" junctions, KRZ for "normal" crossings. Then someone proposed  (KRX), i.e. drop the "Z" and add "X". Following this logic I come to ABY for "Y"-style junctions ... just an idea ... a×p_de^Hello! 22:09, 27 March 2013 (UTC)

Well, if you really want to be correct, it should be ABD: ABzweigstelle in der Form auf einem Delta (but only because it would be too difficult to enter ABΔ from a non-Greek keyboard). Useddenim (talk) 00:14, 28 March 2013 (UTC)

First of all: Don't "correct" my signature! (Sorry, was just testing while hunting for Δ & forgot to change it back. Useddenim (talk) 11:10, 28 March 2013 (UTC))

Apart from that I guess you already realized that this was no serious idea ... a×p_de^Hello! 10:40, 28 March 2013 (UTC)

That someone was me who proposed KRX in August 2011 and made the change in January 2012 (no hasty actions, there!).
The WYE root was introduced and discussed at User talk:Circeus/BSicon renaming/Simple junctions in September–October 2011. As I said then, “I don't see any need to use a German abbreviation (especially one that google won't translate) when there is a three-letter English word that describes the feature perfectly. (Not to mention that the letter Y in the middle of the name makes it completely intuitive.)” There is no need to change it now. Useddenim (talk) 23:37, 27 March 2013 (UTC)

IMHO the  (KRX) is ok, the "KR" is still a good abbreviation for Kreuzung/crossing, and the "X" makes it intuitive which style is meant :)

On the other hand, "AB" is not a good abbreviation for Abzweig/junction, even the attached "Y" won't make it intuitive :(

So I'm ok with "WYE", I would've articulated my opposition otherwise. Btw. the content of User talk:Circeus/BSicon renaming should be moved here as well! a×p_de^Hello! 10:40, 28 March 2013 (UTC)

Why "+" in the WYE names?

(ABZgr+r) →   (WYEr+fg)  (ABZgl+l) →   (WYEl+fg)  (ABZqlr) →   (WYEg+lr)  (ABZq+lr) →   (WYEf+lr)

It's necessary for the orthagonal icons, and then for consistency with the diagonals. Useddenim (talk) 00:05, 28 March 2013 (UTC)

You still haven't brought up an explanation for the "+"!

Actually we have two meanings for it:

1. Originally the compound as in  (GRENZE+BRÜCKE) which has become mostly obsolete outside :de:wp: due to the overlay ability.
2. A bit later introduced to get rid of "lg" and "rg" to enhance junctions. At that time "-" was already used for parallel line icons so I choosed the "+" ... And we made use of it when changing   (ÜWo+l) to  (STR+1).

In the latter meaning "+" stands for "track is coming from {1|l|2|3|r|4}"! Now tell me from where does the track come in "WYEr+fg"? From forward (which is implicit if not noted else) and from the opposite direction (that's where we usually travel to!) ?? a×p_de^Hello! 10:40, 28 March 2013 (UTC)

Apparently it just followed on from the ABZ icons and no one suggested dropping its use. Useddenim (talk) 11:10, 28 March 2013 (UTC)

(Undent) The + in WYE came about, I believe, because when we worked out WYE as "tripoint" icons where all three directions would be explicitly indicated, It felt better to have an "origination" point (in this case, whichever is placed symmetrically to the other two) than to have three direction suffixes without the +. I dunno about you, but  (WYEglr) makes me... squirmy. At least "wye to top from left and right" is a pretty straightforward explanation if someone asks, but we ARE discussing new schemes, so I'm sure UsedDenim is all ears if you got a better idea (especially if that idea deals well with "nonsymmetric" WYE icons, which are bound to arise eventually). Circeus (talk) 15:31, 28 March 2013 (UTC)

It took some thinking, but It's unlikely that something like   would come up very often. And IMHO, keeping the + in the name helps to orient the icon. Useddenim (talk) 18:54, 28 March 2013 (UTC)

You just moved  (ABZ3+l+4) to "ABZl+34" with a link to this talk, but I can't find any line justifying this.

"ABZl+34" translates to: running to left from 3rd and 4th quadrant. But BSicons aren't build this way, we always look in direction of travel (apart of some "lg" and "rg" suffixes). And when looking from top to buttom, first track runs "to left from 4th quadrant", the second one "to 3rd quadrant from left". When we put this into a sequence it is "running to 3rd quadrant from left/to left from 4th quadrant" ergo "ABZ3+l+4"!

Btw,  (WYEl+34) should be "WYE3+l+4" as well. a×p_de^Hello! 16:25, 15 May 2013 (UTC)

No, no, no, no, NO!  You always read icons "in [the] direction of travel"—i.e. from top to bottom—but if you scroll up just a couple of paragraphs you will see:

It [is] felt better to have an "origination" point (in this case, whichever is placed symmetrically to the other two) than to have three direction suffixes …

Frankly, I don't understand why you continue to unilaterally change the names of icons to fit your arbitrary notion of "correctness", especially when there is a general consensus that differs. (Viz.   (ABZ2) vs.  (ABZg2).) However, in this case your logic is particulary. well, illogical:  (ABZ23),  (ABZ+14),  (ABZr+12), and then the only icon with two "+" in its name! (And at risk of causing you to run further amok, you seemingly turn a blind eye to the  (uABZl+34) versions.) I think everyone (else) agrees with the notion of it being easier to understand an icon if there is some symmetry to the name, instead of having to laboriously track its path in one's mind's eye. Useddenim (talk) 19:49, 15 May 2013 (UTC)

1. It's been a rule long before you started to work on BSicons, that they are seen in direction of travel, that's a fact!
2. Following your argument, BSicon names aren't unambiguous anymore! "ABZl+34" would be the same as "ABZ34+l" - but names should be unambiguous, that's why we're struggling so hard to find unique names!
3. Unique file names can only exist, if there is an unambiguous view on all icons. Things getting very complicated, if you got to (quarter-)turn the icons in your mind first to find the correct name!
4. And of course you're right, the other icons should be renamed the same way, but I try to convince you first, that my approach is the correct one. a×p_de^Hello! 19:21, 16 May 2013 (UTC)

1. Sure it is, just like only Bilderkatalog-approved icons can be used at de:WP.
2.  (xENDEl) ⇔   (ENDExl)? Yes, names should be unambiguous. (And unique.)
3. They should also be readily understandable. The illogical conclusion to your argument is the names that DePiep comes up with: (  (DIRECTIONyellow180deg),  (uPANAMACANAL-BASINSr) etc.) If everyone else can agree that WYE and ÜWABZ icons are understandable and can be named based on their symmertical axis, why do you have a problem with that? (And besides, all the q icons are named on the premise that “you [have] to (quarter-)turn the icons in your mind first to find the correct name!”
4. No, they shouldn't be renamed, they should be named consistently. And the correct approach is to discuss first, then rename, not vice versa! Useddenim (talk) 14:24, 17 May 2013 (UTC)

ABZ3+l+4

moved from File talk:BSicon ABZl+34.svg

The name "ABZ3+l+4" means "tracks running to 3 from/to left from 4".

The other proposes name "ABZl3+4" should look like  (STR3+4) +  (STRl+4) =  (STR3+4※ STRl+4) !

Regards a×p_de^Hello! 20:47, 15 April 2013 (UTC)

---

Yet another reason why this is a very bad icon name:

Original ABZ3

 (ABZ3lf)  (ABZ3lg)  (ABZ3rf)  (ABZ3rg)

Newer ABZ3

(all with an origin in corner 3)

 (ABZ3+1g)  (ABZ3+1gf)  (ABZ3+1l)  (ABZ3+1lr)   (ABZ3+r)

Axpde’s unique ABZ3

(which is unrelated to anything else)

 (ABZ3+l+4)

Useddenim (talk) 18:29, 30 June 2013 (UTC)

Well, having both a digit as the last character of a 4-character ID and digits as sufixes appended after 3-letter IDs is asking for trouble. This "ABZ3" thing, however, should have been renamed long ago. For me the solution is simple:

•  (ABZ3lf) → ABZ+lq — or ABZqlf or ABZq+f or ABZq+e; or even a name without a "q", if 3 points and a direction is indicated, such as ABZl+r+f.
• These seem to be all consistent, predictable, and productive:  (ABZ3+1g)  (ABZ3+1l)  (ABZ3+1lr)   (ABZ3+r).
• Things like  (ABZ3+1gf) (and like  (ABZgl+r), too…) would probably be better named with a new ID ^{(say, AZZ)}, for icons showing two opposite branchings from the same main line; however this practice of ABZ plus several origin/destination points doesn’t break the system, per se, only muddles it.
•  (ABZ3+l+4) seems to be both over-specified (3 points) and ambiguous (which connects to which?) or even misleading (goes to 3 from left and 4?, no, that would be  …); I think simpler   (ABZ+3+4q) would be clearer, provided that the double-branch “fork” that looks like   is named with a separate new ID ^{(say, GBL)}. A more traditional approach would be   (ABZl+3+4), indicating one destination (left) and two origins (3 and 4).

And, yes, I’m matter-of-factly suggesting two new IDs to split a major existing one, which also would affect the KRW ankd maybe the kABZ families. Bold, I know. Let me just say that WYE was dettached from ABZ, also, and it was a good move, if also bold. -- Tuválkin ✉ 20:17, 30 June 2013 (UTC)

Regarding  (ABZ3+l+4) - that resembles  (kABZl+l) except for where the lines meet the left corners. So maybe it could include l+l --Redrose64 (talk; at English Wikipedia) 23:32, 30 June 2013 (UTC)

Well, I thought  (ABZl+34) was perfectly clear, but then someone decreed otherwise… Useddenim (talk) 03:57, 1 July 2013 (UTC)

---

At my request, the file has been moved back to  (ABZl+34). YLSS (talk) 15:20, 18 July 2014 (UTC)

Diagonal + curve

moved from User talk:YLSS#What are you doing?!

Compare:

		x	e	ex
ABZrg
hABZrg
ABZg+1
hABZg+1
ABZ3+1g
hABZ3+1g
AB3+1l
hABZ3+1l
hkABZgl

• eABZ = branch out of use
• xABZ = through out of use

Useddenim (talk) 18:25, 15 July 2014 (UTC)

The thing is, we don't have a line through here, only a diagonal line and a curve. And which of them takes precedence? That's a tricky question. If I recall correctly, when I first uploaded any such icons my logic was like this: the curve  (STR+1) has at least one "primary" point ("front"), while the line  (STR3+1) only touches the corners, so may be regarded as subsidiary. (Probably I was even drawn to such a reasoning by something that I read somewhere, but I don't remember now.) Compare the look of the fourth and the fifth rows in the table above. A debatable decision, yes, but I've already uploaded 7 such icons (the red ones), and nobody raised a question (since October!). And since they are all in use, it was simpler to rename the new icons. I hope we agree that consistency goes first? YLSS (talk) 19:30, 15 July 2014 (UTC)

I understand the logic of designating the line that crosses the middle of the icon edge orthogonally as “primary”; and maybe it’s translation problem to/from English, but “through” (as in “straight through”) implies continuous, without deviation. But more importantly (and sticking to the KISS principle),  (uxABZ1+3f) is simply  (ueABZg2) rotated 45°, so I can’t see the logic in switching e and x solely because of a slight reorientation.

As to why I didn’t fix your mis-named uploads: I have to admit that since other editors seem have taken it upon themselves to keep the catalog pages up-to-date with new uploads, I haven’t been watching AlgaeGraphix’ pages as closely as I used to.

And yes, consistency is necessary, but we have to agree on what we’re being consistent about. Useddenim (talk) 01:49, 16 July 2014 (UTC)

---

moved from User talk:Useddenim#Please stop!

Before making any new uploads and duplicating you own (  (uxhABZ3+1l) vs.   (uehABZ3+1l temp)). Let's settle the matter first. Don't create even more chaos. YLSS (talk) 21:19, 15 July 2014 (UTC)

geometry issues moved to Talk:BSicon/Icon geometry and SVG code neatness/Formations#Diagonal + curve

---

Honestly speaking, I'm really upset by your behaviour here, Useddenim! First, quite provocatively uploading the red icons (e.g.  (ehABZ3+1l)), when you have already noticed that I'm not happy with your logic; secondly, brutally undoing my moves by simple overwriting when the discussion was already in progress (e.g.  (uxhABZ3+1g)), so that their reversal becomes even more complicated (but not impossible); and thirdly, messing with other files (e.g.  (ueABZ3+1g)) so that they start corresponding to your logic. And if presented some 100-pound arguments here that justified my POW? Really, I didn't expect that from you.

That said, I am not against switching to your naming scheme, as purely algorithmically they are equally possible. But if so, I would expect you to deal with the mess that now can be seen here, since it was you who started uploading files without checking if they agree with the existent patterns. (And don't start telling that they were wrong, 'cause they were not!) That includes renaming and replacing  (xABZ1+3f) and similar, and please do that properly, by a three-step renaming, because as I said, I do not like seeing files overwritten by different files (except for improvements). I expect this as a gesture of goodwill from someone who has shown such bad discourse practice. Otherwise, I would regard myself in full right to rename your files to follow the pattern that I choose originally: just because I regard them as equally justifiable, and because the newer files are not in use as yet. YLSS (talk) 06:40, 16 July 2014 (UTC)

And I was upset with you switching my uploads ( (uetABZ2+4g) &  (uxtABZ2+4g)) without bothering to say anything, and if I recall correctly, you pretty much overrode anyone’s objections to the l/r →‎ fq/gq conversion, but two (or three or four…) wrongs don’t make a right. For the record, I only changed the unused icons, leaving the aforementioned red ones alone pending resolution of this issue. (And for what it’s worth, something’s not caching correctly with  (uxABZ2+4g), as the image displayed by WP is not the local image on my computer.)

I’m OK with me fixing up the inconsistencies at WP:RDT/BS-ÜWjunctions, and—in fact—icons such as  (ABZ1+fr),  (ABZ23) et al.—i.e. without a clear primary route—should follow the pattern of ABZ1+xfr & ABZ1+fxr and ABZx23 & ABZ2x3 rather than eABZ1+fr & xABZ1+fr and eABZ23 & xABZ23.

As I see it, calling the x=250 or y=250 point “primary” is a weaker argument than either “a straight line takes precedence over a curved one” or “the established pattern used with icons not rotated 45° should be followed”.

The discussion you referred to (“…something that I read somewhere…”) sounds vaguely familiar, possibly something that Axpde said, either when the ÜW set (now STR+1-STR2-STR3-STR+4) was first being expanded, or during the ABZ2/ABZg2 renaming debate (but I couldn’t locate it either…). Useddenim (talk) 20:32, 16 July 2014 (UTC)

I switched  (uetABZ2+4g) &  (uxtABZ2+4g) because I thought that was a lapse of attention on your side (which happens to all of us). I did not expect that they were backed up by such a adamant principles...

Yes, I think the best way out of this would be to settle on  (xABZ1+3f) →  (ABZ1+3xf) and   (xhABZ3+1l) →   (hABZ3+x1l). After all, when I though about how to rename ÜWo+r-eABZrg, I decided on   (ABZf+4xl) just because I was not sure which line to consider as primary. So it would have been more prudent of me to do the same with  (ABZ1+3f) variations. So... do we settle on this? YLSS (talk) 10:24, 17 July 2014 (UTC)

P.S. The new {{WP:RDT/BSa7x}} will be handy for this. YLSS (talk) 11:39, 17 July 2014 (UTC)

Well… those were the names I was planning to use during the renaming (instead of temp), but I still prefer the simplicity of xABZ for the straight line and eABZ for the curved one. And after tinking about the whole issue some more, I seem to recall that were was some discussion, and that it was decided that the straight line was the primary route and the branch was secondary (which is why the icons are called  (ABZ1+3f) rather than ABZ1+f3 etc.). Useddenim (talk) 12:30, 17 July 2014 (UTC)

Evidently there is no "simplicity" in "e/x", since we differ in opinion on which line should be primary... YLSS (talk) 18:28, 17 July 2014 (UTC)

So then it's  (eABZ3+1g) &  (xABZ3+1g) (per your comment)? Useddenim (talk) 20:47, 17 July 2014 (UTC)

? I don't get your logic. What's the connection? YLSS (talk) 21:48, 17 July 2014 (UTC)

“But feel free to choose the name you want, as the inventor of the series.” I believe I was the first to upload diagonal ABZ icons… Useddenim (talk) 00:09, 18 July 2014 (UTC)
 (ABZf+1l) on 8 April 2012 (admittedly experimental and obscure); or  (utABZ1+3r) on 2 November 2012. Useddenim (talk) 00:18, 18 July 2014 (UTC)

Heh, but that I can claim to be the inventor of the diagonal eABZ/xABZ icons, the ones that we discuss here ;) But let's not go down that road... YLSS (talk) 05:44, 18 July 2014 (UTC)

If I may add something that may add to this. I have always looked at e and x situation (in junctions) as referring to straight and divergent tracks. In this instance, straight (or "through," if you prefer) to me signifies the diagonal track, not the ones originating/ending at g, l, f, or r (as these tracks join or diverge the diagonal). Lost on Belmont (talk) 13:07, 18 July 2014 (UTC)

I agree with Lost on Belmont here. The straight (or least bent) line is the “main” one. Seems to be a simple issue, anyway. -- Tuválkin ✉ 14:15, 18 July 2014 (UTC)

You chose a very good time, guys, to join our argy-bargy here, because I have just started moving the files to names with x "presuffix" (as described above) before noticing your responses! (Also, I have renamed eABZl+34 →  (ABZl+3x4) etc. following Useddenim's suggestion above, especially so because we already had  (ABZ2x3).) Oh well... Since this appears to be another case where some people think this way and some another, like with CONTs and ENDEs, I suggest that the files should be located under the uncontested names, which are the "presuffixed" ones like  (ABZ2x3). I planned to list the redirects from "e"/"x" titles for deletion, but since I appear to be in the minority now, I'll leave them as they are, so that they can be re-targeted as you consider best (as is ueABZ23); but please add them into Category:Icons for railway descriptions/redirect/ex, if so. I have only one question: where have you been before?! Because I didn't upload those 7 icons in a single day, but did that when I needed them, so between October and July they must've appeared in several of AlgaeGraphix's pages (and I know that at least Tuválkin checks them). So please next time that you notice me on the wrong track, please notify me before I become used to it... YLSS (talk) 15:46, 18 July 2014 (UTC)

Since everybody apparently lost interest in this topic, I did all the dirty work by myself (as usual). So now all concerned files are located at "presuffixed" titles, with redirects to them from titles with "e"/"x" as preferred by the majority, i. e.  (xABZ1+3f) =  (ABZ1+fx3). YLSS (talk) 22:44, 28 July 2014 (UTC)