Talk:BSicon/Renaming/SPL

Close parallel across

moved from User talk:Wiebevl

A number of new Parallel Lines icons were posted into this discussion, and I thought you may want to review the names of some of them, particularly:

• (exvSTRqgr-STRqgl) ?→ exvSTRgl-g+l   (exSTRl-STR+l)
•   (exvSTReq) ?→ …?
•   (exvSTRaq) ?→ …?
• (exvABZer) ?→ exvABZg+r-g+r   (exSTR+r-ABZg+r)

Useddenim (talk) 11:08, 18 June 2011 (UTC)

Yes, please do: I made those but was not at all certain about naming. Sorry about that. --Tuvalkin (talk) 18:35, 19 June 2011 (UTC)

I seem to have missed some of this discussion — can someone point me to the missing thread? Useddenim (talk) 14:07, 20 June 2011 (UTC)

Thanks for updating :-) I'll check! Wiebevl (talk) 14:20, 20 June 2011 (UTC)

moved from File talk:BSicon exvBHFl.svg

Oops, I just noticed that   (vBHFl) is very different from this. I chose the name inspired by things like   (uvENDEl) and   (exvCONTl), but seems that here is needed something like vBHFlq…? --Tuvalkin (talk) 08:49, 24 June 2011 (UTC)

Yes,   (exvBHFl) should most certainly be a "q" icon. Useddenim (talk) 15:14, 24 June 2011 (UTC)

Should definetely have a "q"-suffix. But I have no clue how to explain the twin tracks to the right and the single track to the left!

Btw. I don't like the habit to use "l"- or "r"-suffix to indicate a continued station to the left or right, the "l"- and "r"-suffix should always refer to tracks! axp_de^Hello! 18:17, 24 June 2011 (UTC)

"l" & "r" applied only to tracks:

a) not a bad idea, but may be too late to change; and

b) I can't think of a logical alternative other than "<" and ">" — and as appealing as it may be, I don't think it would be a particularly good idea to introduce more special characters into BSicon naming. Useddenim (talk) 18:55, 24 June 2011 (UTC)

But I don't know. Maybe   (vBHF<) would work... Useddenim (talk) 19:05, 24 June 2011 (UTC)

<barbie>BSicon naming is hard!! ^o^</barbie> Okay, guys, I was hoping this is an easy one. Assuming that   (vBHFl) is not going to be renamed (I hope not!), this new one should be either exvBHFlq or exvBHFql. What’s the verdict? --Tuvalkin (talk) 00:18, 25 June 2011 (UTC)

I am not familiar with the overall convention, and maybe this is outdated, but en:WP:RDT/C clearly states that -q "should only be used if the base icon can be rotated in either direction to produce the transverse version." (i.e. if it is symmetrical both vertically and horizontally, hence  (exvBHFq)). This is clearly a rotated  (exvBHFa), and so the result would probably best follow the pattern of normal stations where -r and -l indicate rotatin (e.g.  (KBHFr)), resulting in   (exvBHFar). In general though this is one of very few icons that I can find with transverse transfer stations, so we are treading in unknown territory, but I think the logic is sound. Circeus (talk) 15:42, 21 July 2011 (UTC)

Well, quite the contrary the passage about "q" has been added recently by User:Useddenim to stop people turning an icon by 90° and adding the "q" to the old iconname ... but some icons were turned clockwise, some counterclockwise.

In this case we have an icon that definetely runs across, but with parallel railways to the right. I'd say we should ask User:Wiebevl, he already named alot of BSicons! a×p_de^Hello! 11:12, 22 July 2011 (UTC)

I would strongly advise against icon names containing either < or > because these characters have special meaning (a) to operating systems including UNIX/AIX/LINUX and the MS-DOS/Windows groups and (b) within SGML/XML/HTML/etc. markup. --Redrose64 (talk) 11:28, 22 July 2011 (UTC)

Have found a document which is directly relevant to this: en:Wikipedia:Naming conventions (technical restrictions)#Forbidden characters. --Redrose64 (talk) 10:05, 23 July 2011 (UTC)

Is there a general page for asking about BSicon names? I would like to know about canal swing bridges, particularly those where the bridge carries a railway (i.e.  (umKRZusw) and the two I recently created  (umKRZqusw) and  (uxmKRZqusw)). --Redrose64 (talk) 11:28, 22 July 2011 (UTC)

I suggest Talk:BSicon, ok? a×p_de^Hello! 13:32, 22 July 2011 (UTC)

Thanks, have asked at Talk:BSicon#Icon name query. --Redrose64 (talk) 19:25, 22 July 2011 (UTC)

Re: -q,the problem remains in that it leaves us wondering how to name the mirrored icon to the one at issue here. Circeus (talk) 13:53, 22 July 2011 (UTC)

Is it necessary to use just one character? If two can be permitted, how about "bl" and "br" - the "b" is short for Bahnhof. That is to say, move the existing   (vBHFl) to BSicon vBHFbl.svg --Redrose64 (talk) 10:05, 23 July 2011 (UTC)

  (vBHFl) is not questioned, although I posted somewhere else that the "l" is somewhat irritating, the suffix "l" means "track running to the left". I suggested to call this icon "vBHFR" (station continuing to the right). a×p_de^Hello! 11:52, 23 July 2011 (UTC)

  (exvBHFl) is not the correct name i.m.o. since the 'q' suffix is lacking, using < or > is no option either due the the above mentioned reasons. Therefore I would suggest to name this specific one   (exvBHFqe), reading from left to right the parallel track ends.

Renaming   (vBHFl) to something else is another discussion, I chose this name not because of the orientation of the tracks, but because of the orientation of the station, combined with a matching symbol, either   (vBHFr) or   (dBHFr) it's on the left. Wiebevl (talk) 21:18, 25 July 2011 (UTC)

I agree with Circeus (talk) that the appropriate name for   is (exvBHFar): line & feature out-of-use | parallel lines | STATION | start (of parallel lines) | direction to the right. (I don't believe that anyone has any objection to restricting "q" strictly to bidirectional (or non-directional) transverse icons. Useddenim (talk) 02:04, 26 July 2011 (UTC)

  (exvBHFar) would be the disused version of   (vBHFra) (albeit with the suffixes swapped, something still to agree on) so I still propose   (exvBHFqe). Wiebevl (talk) 13:15, 26 July 2011 (UTC)

What would the horizontally mirrored version be?   (exvBHFeq)?

  (exvBHFqa) imo. Wiebevl (talk) 11:50, 1 August 2011 (UTC)

Although you are used to name those icons, I don't like the "e" or "a" suffixes with horizontal icons! There's no rule for icons "reading from left to right" ...

•  (vSTRa) and  (vSTRe) are ok
•  (exvSTRaq) should be   (exvSTRql)
•  (exvSTReq) should be   (exvSTRqr)

Same with this icon ... a×p_de^Hello! 19:48, 2 August 2011 (UTC)

BTW, I must say I see no reason whatsoever that the disused version of   (vBHFra) cannot be   (exvBHFra). Why should straightforward prefix addition forces a modification of the suffix? It seems logical that the all-disused version of an icon should always be obtainable by straightforward addition of ex-. Creating this convention for across icon verge on the nonsensical AFAIAC. Circeus (talk) 20:12, 2 August 2011 (UTC)

I agree with Circeus that exICON should simply be an out-of-use version of ICON. However, there is still a problem with a×p_de's suggestion of   (exvSTRql) and   (exvSTRqr):

ex = out of use

v = parallel lines

STR = STRetch of track

q = line runs aqross

l/r = to left/right

? = single to double transition

Since we can compound prefixes as needed, there's no apparent reason that the same can't be done with suffixes; so in this case maybe the names should be exvSTRqal and exvSTRqar? Useddenim (talk) 23:29, 2 August 2011 (UTC)

[undent] The a/e suffixes mean "beginning (of double track)" (Anfang) and "end (of double track)" (Ende), so they make perfect sense for distinguishing these icons (though I'm not clear what the correct default horizontal direction is). I'm afraid Axpde's suggestion just does not work at all: -r/l is used for rotation or combined to indicate direction, so combined with horizontal unused double line ( (exvSTRq)), it honestly makes no sense. -a/e works fine here.

I'm willing to go for   (exvBHFeq) to match   (exvSTReq) (but -qe/-qa should be right out IMHO) as an alternative to   (exvBHFar) (note that -r indicate a straigt-up rotation, so that suffix takes the -a of the unrotated icon, the mirrored being   (exvBHFal), but -q presumes a default horizontal direction, which seems to be right-to-left, resulting in a mirror of   (exvBHFaq)). I think either option is fine for renaming this, and unlikely to run into conflict. Circeus (talk) 17:37, 3 August 2011 (UTC)

Conclusion?

I suggest this naming for icons where the parallel lines run across:

1. If the icon has rotational symmetry, then the name is the same as the regular icon, with the addition of the suffix q (denoting across):

(vSTR)	(vSTRq)
(uvBHF)	(uvBHFq)

2. All other icons are named as if they were two regular icons one above the other. Note that there is no v prefix for these icons:

(meKRZ-KRZo)	(emKRZo)  (KRZo)
(uexSTRl-STR+l)	(uexSTRlf)  (uexSTRrg)

3. Unfortunately, this still leaves a problem with the icons that transition from parallel lines on one side to regular on the other (and vice-versa):

 (vSTRl) vs.   (exvSTRaq).

Useddenim (talk) 13:21, 9 September 2011 (UTC)

For the single to double transition, if the preferred direction is designated as "from single to double", then all of those icons then become a variant of
(vSTRar) and (vSTRal)
(or is that (vSTRra) and (vSTRla)?). Useddenim (talk) 04:07, 19 September 2011 (UTC)

IMO directional suffixes should always come last. I personally prefer a system that opposes -e and -a systematically:   (vSTRqa) &   (vSTRqe), but as long as names are regularized (and that   (exvBHFl) is fixed at the same time), I'll be happy. Circeus (talk) 06:44, 19 September 2011 (UTC)

r and l are more intuitive than qa and qe. Useddenim (talk) 11:16, 19 September 2011 (UTC)

Maybe, but they lend themselves too easily to there being multiple possibilities, plus this format causes them to have two distinct meaning: one related to turns and splits, and another related to directionality. The only reason that -q is supposedly less intuitive is that the system never posited a default direction of travel. That's just my grain of salt, I'll live with anything that regularises with that horizontal station icon. Circeus (talk) 16:08, 19 September 2011 (UTC)

Proposed new SPL root

SPL: for the transition between single and double lines; en: SPLit; de: SPaLt

(SPLa)	(SPLl)	(vSPL+la)  (SPLa+l)	(vSPL+ra)  (SPLa+r)	(vSPL+re)  (SPLe+r)
(SPLe)	(SPLr)	(vSPLle)  (SPLel)	(vSPLre)  (SPLer)	(vSPLl+4e)  (SPLe+4)		etc.

Reasoning

1. Icon has a unique track feature, and is not a simple variant of STR.
2. Icon should not have a v prefix because it also connects to a single line.
3. The root rationalizes and simplifies naming, especially for parallel transverse icons (  (STR+r-ABZg+r),  (vSTRl), vs.   (exvSTRaq) etc.)
4. Allows for correct naming of  (tvSTRa) and   (vTUNNELa) ( (tvSPLa) &  (tvSTRa)) and ends.
5. Removes 1/4 shift icons from the BS2 group (which should only be for lines that start or end at the edge of the icon) (replaced by   (SHI1)):

(vBS2r)	(vBS2l)	(vBS2+r)	(vBS2+l)
(vSPLa-)  (SHI1r)	(v-SPLa)  (SHI1l)	(vSPLe-)  (SHI1+r)	(v-SPLe)  (SHI1+l)

Useddenim (talk) 14:24, 28 November 2011 (UTC)

Circeus had suggested that there may be a problem with the naming of the icons where there is the start/end of both double lines and another feature, but this could be adequately covered with a double suffix, similar to icons where there is both directionality and displacement (e.g.   (uINTll) or   (WSLer)):
 (vSTRa) +    (hNULe) =   (hSPLae)
 (vSTRa) +    (hNULa) =   (hSPLaa)
 (vSTRe) +    (hNULa) =   (hSPLea)
Useddenim (talk) 05:20, 29 November 2011 (UTC)

1. Just because it is a "unique feature" doesn't mean it isn't handled fine for the time being (and it's not that unique a case: I'm easily halving the number of station roots in my proposal!).
2. On the contrary, if parallel lines enter the icon anywhere, there should be a v- (that's my key criterion for the various curve things)
3. These are badly named in original system! I deal with transverse icon with a broader, more efficient solution. (plus   (STR+r-ABZg+r), as you prove yourself, is no transverse icon! It's a curved   (vSTRe)!)
4. You prove introduction of a different problem (what is  (tvSTRa)+ (TUNNELa)?). Currently these various problems are solved by NEITHER vSTRa/e OR SPLa/e (your "offering" has double suffixes and icons differentiated by arbitrary changes in the order of suffixes, both of which are also reasons why I wanted to propose a broader rename), and that alone is good enough a reason for me to recommend waiting until someone can think of a solution.

   Unless my logic is faulty, I would assume that the first suffix—i.e. the one immediately adjacent to the root—would apply to the root, and the other suffix would apply to the prefix, "bracketing" the root. So to use the example above, the answer is   (tvSPLaa), where vSPLa is the single-to-double transition, and tvSPLaa denotes the beginning of a tunnel.

5. I introduce a system capable of handling a single track going from any currently in use point on the side to any other (there are currently 18 total of them), and makes all of them STRs (which the highly counterintuitive BS2s are FAR closer to than to the "unique featured" of splits!), not to mention that one of my priority is to eliminate the dash convention for single tracks.

So in the end I don't like SPL because a) it hardly addresses the problems that STRa/e already has and b) fails to actively shorten the icon names (as I've noted, I believe dropping the v- in these is a bad idea), making the gain nil. The root may be a marginal improvement in intuitiveness, but carrying over the problems of STRa/e makes me say the "fix" isn't that useful.

Circeus (talk) 06:23, 29 November 2011 (UTC)

I'm still not entirely happy about the whole "unneeded name replacement", bu I have found a solution to the suffix conundrum:

•  (vSTRa)   (vSPLf)
•  (vSTRe)   (vSPLg)
•  (vSTRl)   (vSPLgq)
•  (vSTRr)   (vSPLfq)

That is, instead of the meaning "beginning/end of double line" we have the meaning "split forward/backward". Circeus (talk) 15:42, 11 December 2011 (UTC)

I like f/g instead of a and e, but I still prefer l and r over q…f/g, which then gives us

•   (vSPLf)
•   (vSPLg)
•   (vSPLl)
•   (vSPLr)
•     (vSPLf-)
•   (v-SPLf)

etc. Useddenim (talk) 18:40, 11 December 2011 (UTC)

And isn't putting q ahead of the root introducing yet another variation? Useddenim (talk) 18:42, 11 December 2011 (UTC)

Huge brainfart on my end. And I STILL believe that BS2 icons ought to be merged into the STR root. Circeus (talk) 20:01, 11 December 2011 (UTC)

No problem with BS2 (½ & ¼ shift) → STR; but what do you propose to use to indicate them? (we already are using numbers to denote the different corners.) Useddenim (talk) 20:25, 11 December 2011 (UTC)

A further thought: these already exist:
 (STRf)  (STRg)  (STRr) &  (STRl);
so the v versions should be something like  (vNULf※ vSTR)  etc. Useddenim (talk) 13:23, 12 December 2011 (UTC)

These icons were one of the horrible decisions from early icon-naming. I'm not sure yet, but I may introduce a new root for those. They are essentially taking up names too useful to be hogged by such rarely used icons. Given that I have never seen them actually used with other roots, I don't think they are all that liable to cause problem with vSPLg. Circeus (talk) 14:50, 12 December 2011 (UTC)

We have already introduced  (BS2lr), so why not using the same root for parallel icons? a×p_de^Hello! 15:19, 16 December 2011 (UTC)

Because BS2 is for shift from the center line to the side (more generally the BS2 icons are originally meant to be "column linkers", restoring alignments where one line has a different number of standard-width icon). The BS2 set is just another bad choice of root (BS="Bahnstrecke", that is "railway"!) and the entire set has consistently counterintuitive names, many of which takes the room of what should by all mean be something else. Moving a longstanding well-understood name into that can of worms is, I'm sorry to say, possibly the worst solution offered so far in this discussion. Circeus (talk) 22:44, 16 December 2011 (UTC)

I know, the root BS2 is maybe not the best choice, but honestly I see no better one. BS2 means "split into two railways", and this works quite well. Ok, it was me who proposed "2" to indicate tracks running to the lower right corner which limits this root ID somehow. On the other hand: Do we really need *all* combinations of thinkable tracks?hv-STR+1

I just saw User:Tuvalkin's creations  (hv-STR3o2+4),  (hvSTR3-),  (hv-STR+1) and  (hvSTR-L+1-)^[renamed], which are pretty much of no general use ... a×p_de^Hello! 09:30, 17 December 2011 (UTC)

Unified naming:

Present	(vSTRe)	(vBS2+r)	(vBS2+l)
Proposed	vSPLg	vSPLg-	v-SPLg
Present	(BS2+lr)	(BS2+r)	(BS2+l)
Proposed	SPL+14	SPL+4	SPL+1
Present	(vSTRalr)	(utvBS2+r-)	(utv-BS2+l)
Proposed	vSPL+14	utvSPL+4	utvSPL+1

etc. Useddenim (talk) 18:30, 17 December 2011 (UTC)

As far as I'm concerned, SPL should NOT be used for the half-width shifts (both because they cannot be considered splits and because I vigorously disapprove of the whole dash concept). Also I am not convinced that merging   (vSTRalr) into the new SPL is such a good idea. Going back to axpde's proposal, I think I can better articulate my thoughts now that I see UsedDenim's proposal: Axpde's look is backward! BS2 has a very clear "feel": column shifts, most of it to the corner. The SPL root is clearly NOT a column shift, and these things are conceptually different despite the initial similarity (this is the same why we have  (KRWlr) and  (ABZlr) into separate roots). As such I think ANYTHING that uses the BS2cX corners should be left out of a proposed SPL root. That subset of BS2 forms too coherent a family to be shoehorned into two different roots! Circeus (talk) 06:19, 18 December 2011 (UTC)

Alternate scheme:

Present	(BS2+lr)	(BS2+r)	(BS2+l)
Proposed
Present	(vSTRalr)	(utvBS2+r-)	(utv-BS2+l)
Proposed	(BS4+lr)	(tBS4+r)	(tBS4+l)

Transverse versions would use the f/g suffix instead of l/r. Useddenim (talk) 13:45, 18 December 2011 (UTC)

1/2 shift vs. 1/4 shift

The BS2 root covers icons that shift the line over one-half the icon width (250 pixels):  (uBS2l),   (uvSTRel).
However, there are a number of parallel straight icons that only make a quarter-width (125 pixels) shift:   (vSTRel),   (utvBS2+r-),   (utvBS2+r),   (utvBS2l). Some method needs to be devised to address these.

1/2 shift (BS2)

Four combinations:
Description		Existing		Proposed
right edge to center		(BS2+r)	(tBS2lf)	→  (BS2+r)	Existing names are OK, but "old-style" suffixes need to be standardized.
center to right edge		(BS2r)	(tBS2rg)	→  (BS2r)
left edge to center		(BS2+l)	(tBS2rf)	→  (BS2+l)
center to left edge		(BS2l)	(tBS2lg)	→  (BS2+l)

The right-to-left lines (and reverse) can be ignored for now, as they seem to be named adequately:
  (vSTRr2l)   (vSTRl2r)   (vSTRr2lq)   (vSTRl2rq).

  (uvSTRel) is not a column shift, because the entering and exiting are perfectly aligned, it's just   (uvSTRr2l) that happens to be using an ACTUAL v-type line (these will be part of a coherent proposal for single-line v- icons, as they currently violate one of my basic rule by containing multi-character suffixes making them no better than the legacy -lf/g stuff). The l2r/r2l icons probably ought to acquire a dash to free the space for   (uvSTRel) and its companion. In any case, as long as the "edge" icons keep to a BS2 root, I don,t think anything specifically needs to be done. Circeus (talk) 05:14, 10 September 2011 (UTC)

1/4 shift (BS4)

Eight combinations:
Description		Existing	Proposed
right edge to right line		(utvBS2+r-)	→   (utBS4+r)
right line to right edge		(n/a)	→   (utBS4r)
right line to center		(utvBS2+r)	→   (utvSTRr2c)  (utvSTRe-)  (utvSPLe-)
center to right line		(utvBS2r)	→   (utvSTRc2r)  (utvSTRa-)  (utvSPLa-)
center to left line		(utvBS2l)	→   (utvSTRc2l)  (utv-STRa)  (utv-SPLa)
left line to center		(utvBS2+l)	→   (utvSTRl2c)  (utv-STRe)  (utv-SPLe) (Useddenim (talk) 05:26, 29 November 2011 (UTC))
left line to left edge		(n/a)	→   (utBS4l)
left edge to left line		(utv-BS2+l)	→   (utBS4+l)

Incomplete, and also doesn't really follow the naming rules very closely. 13:57, 10 September 2011 (UTC)

Any suggestions anyone? Useddenim (talk) 04:04, 10 September 2011 (UTC)

Lines to center are single line vSTRa/e, which should be doable with a dash. The four others are BS4 forms of the BS2 icons. Combine these sets by hyphenating them together like ¾ of the v- icons already are and voilà. Circeus (talk) 05:14, 10 September 2011 (UTC)

  (utvBS2+r) →  (utvSPLe-) and so on. Personally I LIKE the v- + -a/e system (which I intend to apply in my proposal, including for -q icons), but if you don't, then I got nothing. My idea is that only icon that straddle the line should have the BS# roots. It's my take on it, but I do think it frees up things. Circeus (talk) 18:56, 10 September 2011 (UTC)

I'm not sure how to go on these, but would like some consistency between BS2 and BS4, and a naming scheme that would also be applicable to transverse parallel icons. Useddenim (talk) 20:20, 10 September 2011 (UTC)

Of course if we define BS2/4 icons differently, it doesn't help. I firmly believe it is more efficient to treat under the BS root ONLY icons that runs at the edge, and treats the other within a coherent global naming scheme. Circeus (talk) 05:11, 12 September 2011 (UTC)

More bad names…

Unfortunately, I think I was the one who did theses. :(

current old name	should be new name	single version
(vSTR+r4-)	(vSTR+4-)	(STR+4)
(vSTRr3-)	(vSTR3-)	(STR3)
(vSTR+l1-)	(vSTR+1-)	(STR+1)
(v-STR+l1)	(v-STR+1)	(STR+1)
(v-STRl2)	(v-STR2)	(STR2)
(vSTRl4-)	(STRl+4-)	(STRl+4)
(vSTRr1-)	(STRr+1-)	(STRr+1)
(vSTR+l3-)	(STR3+l-)	(STR3+l)
(v-STR+l3)	(-STR3+l)	(STR3+l)
(v-STR2+r)	(-STR2+r)	(STR2+r)
(v-STR+r2)	(-STR2+r)	(STR2+r)

Useddenim (talk) 11:45, 27 January 2012 (UTC)

Some of these renames left redlinks and busted diagrams I found only now, 3 weeks later (like this one). This is not good. -- Tuválkin ✉ 13:06, 18 February 2012 (UTC)

And here, too. Really not good: Renamings should be made without leaving a trail of defective diagrams. -- Tuválkin ✉ 21:51, 18 February 2012 (UTC)

The problem's with your browser. (Try purging the cache.) Check Global usage for "File:BSicon whatever.svg" and you'll find the message "is not used on other wikis". Useddenim (talk) 23:51, 18 February 2012 (UTC)

I'm talking about red links in these diagrams (at least these two, there may be more), caused by renamings that were done without checking and fixing. It is not about file redirects, browser cache, and blue links for icons not loaded in the diagram. It is about what needs to be fixed by editing the wiki code behind these diagrams, search for filenames of deleted icons, and replacing them with the new names — which should have been made before the file with the old name was deleted. -- Tuválkin ✉ 00:09, 19 February 2012 (UTC)

The only red links I can see are P.^ça Quebedo, pt:Movauto, Praias do Sado & Mourisca-Sado in pt:Predefinição:CP - Praias do Sado, and Formoselha, Montemor-o-Velho & Vila Nova de Anços in pt:Estação Ferroviária de Alfarelos, which I had no part of creating. If you actually want to be helpful, then state which icon is not displaying properly, instead of making spurious claims of malfeasance and bad faith editing on my part. Useddenim (talk) 13:21, 19 February 2012 (UTC)

You could not see the red links messing up the diagrams because I fixed them had first. Here's the relevant history items:

• {{:pt:CP_-_Praias_do_Sado}} hist diff
• [[:pt:Estação_Ferroviária_de_Alfarelos]] hist diff

The deleted icons affecting these diagrams were vSTRr1-, v-STR+r2, v-STR2+r, and v-STR+l1. -- Tuválkin ✉ 19:24, 19 February 2012 (UTC)

So, let me summarize, just to make sure I understand the situation: Last year you downloaded a "dump" list of some 840 redirects that likely encompassed thousands—if not tens of thousands—of uses. In the meantime you've done ??? about eliminating any of them, but complain when I missed two diagrams out of several hundred edits ? Useddenim (talk) 00:25, 21 February 2012 (UTC)

Well, seems that’s not more bad names, but more good names to replace existing bad ones. :-) -- Tuválkin ✉ 17:33, 27 January 2012 (UTC) taking it back, as I think now these now names are not 100% perfect. Tuválkin ✉ 22:20, 29 January 2012

If

(v-STR+l1)	(v-STR+1)	(STR+1)
(v-STRl2)	(v-STR2)	(STR2)

then

(vSTR+l3-)	(STR3+l-)	(STR3+l)
(v-STR2+r)	(STR2+r-)	(STR2+r)

Circeus (talk) 21:00, 27 January 2012 (UTC)

Ah, apparently I reversed those two. Useddenim (talk) 04:54, 28 January 2012 (UTC)

 Done. Useddenim (talk) 18:37, 29 January 2012 (UTC)

---

As detailed in the next section, exemplified by  (v-STR2), I don’t think this renaming is correct. All these are either the left or right half of parallel lines, yet converging to (or coming from) one single corner at the edge of the icon. Meaning that were both present on the icon, it would be something akin to  (vSTRa) or  (vSTRe), not  (vSTR) — therefore their halves need to include an "a" or "e" in the name. -- Tuválkin ✉ 22:41, 29 January 2012 (UTC)

Nope. A complex icon's name should relate to a simpler version of the same (as detailed in the original chart above), not to some what-if? potential use. Besides, consider this combination:  (v-STR2) +  (STR2+r-)  (v-STR2※ STR2+r-)  clearly much more akin to ABZ than   (SPL). Useddenim (talk) 23:57, 29 January 2012 (UTC)

Having said that, I have no problem with naming  (vSTR2-※ v-STR2)  as   (vSPL2e), as a derivative of (vSPLe). Useddenim (talk) 00:04, 30 January 2012 (UTC)

And riddle me this, Batman:   ? Useddenim (talk) 00:10, 30 January 2012 (UTC)

So we agree on the fundamentals, that  (vSTR2-※ v-STR2)  in every way comparable to , and likewise for all these doubles converging to a single corner. Now you just need to notice that that the table above is filled with halves of these doubles and you’ll swap your "nopes" to "yopes". It is not a what-if case: All those icons need an "a" or an "e" in their names so that they don’t use up a name that is needed for half double lines going off to a corner but not convergent to a single line, as detailed in the next section. -- Tuválkin ✉ 20:39, 30 January 2012 (UTC)

I mean specifically that (what is now named)   (v-ÜWol) was originally uploaded by me as  (v-STR2) because it is its logical name, given  (STR2),  (vSTR), and  (v-STR) (like you said, a «complex icon's name should relate to a simpler version of the same», and that’s what I did); it is not a far-fetched what-if case created to ruin your naming scheme. When I created and uploaded that icon, I simply needed it for a diagram, as innocent as that. All the names above just need an "a" or an "e" and all will be okay. -- Tuválkin ✉ 20:51, 30 January 2012 (UTC)

As for the unrelated examples you gave —   is akin to  (vSTR-STRl), and indeed as you said   is an ABZ (if at all worth to be created and named as single icon, given its random topology) — so what? These two are not what I’m talking about.

-- Tuválkin ✉ 20:39, 30 January 2012 (UTC)

[undent] I think tuvalkin makes a VERY valid point. We need to manage a way to separate   (vÜWol) (AKA  (vSTR2)),   (v-ÜWol) (which is logically  (v-STR2)) and  (v-STR2) (AKA the former   (v-STRl2)). Well, at least now we know where those name came from! IMHO, I think these are a GREAT example of icons that beg to use my 1/2 numbering prefix scheme (IF we use within a system where the dash system is also in use: otherwise the problem resurfaces). Circeus (talk) 22:55, 31 January 2012 (UTC)

So, going back to Proposed new SPL root above, are people now willing to accept a renaming of the single-double/double-to-single transitions, to at least remove some potentially dubiously named icons from the vSTR family? Useddenim (talk) 23:55, 31 January 2012 (UTC)

I completely fail to see what sort of connection these have to any possible SPL icon. Circeus (talk) 00:50, 1 February 2012 (UTC)

Unrename vk-STR2, please?

Legend vSTR + uSTR v-STR + uSTR File:BSicon v-ÜWol.svg v-STR2 + uSTR2* File:BSicon v-ÜWol.svg v-STR2 + exkABlf* File:BSicon v-ÜWol.svg v-STR2 + exkABZgl* (*) Renamings: v-str2 → vk-str2, 2012.01.27 vk-str2 → v-ÜWol, 2012.01.28

Legend STR ≡ STR2 (just to illustrate how it compares) vÜWol vSTR ≡ vÜWol (vÜWol should be vSTR2) vSTRe ≡ vSTR2- + v-STR2 (vSTR2- + v-STR2 should be vSTR2e?)

Concerning this:

27 January 2012 11:31 . . Useddenim (talk | contribs)‎ moved File:BSicon v-STR2.svg to File:BSicon vk-STR2.svg

However, I’m sure this (now renamed   (vk-STR2)) has nothing to do with the k series. This is the left-side-only variant of a double track that follows the same exact path as  (STR2), hence the name I chose originally. It is perfectly parallel to it, and not to the Coumpound Junctions’ icons. See it here → overlapping with  (uSTR2), parallel, and with  (exkABZgl) and  (exkABlf), not parallel (colored for contrast). Unless I’m missing something, please “unrename” this file back to  (v-STR2).-- Tuválkin ✉ 18:33, 28 January 2012 (UTC)

I renamed the file to   (vk-STR2) because  (v-STR2) is/should be the correct name for   (v-STRl2). (See the previous discussion topic.) Having said that, though, 1) I admit to selecting the new name too hastily, and 2) a naming scheme needs to be developed to describe half of a double line exiting a corner. Useddenim (talk) 19:08, 28 January 2012 (UTC)

Meanwhile I saw that you needed this name for  (v-STR2) — however, note that in this case we have a double line (or any half of it, as in  (vSTR2-)) converging to a single point, akin to  (vSTRe) (or SPL, but I’m not going there). Therefore what you now named  (v-STR2) should be   (v-STRe2), being the left side half of (a future)   (vSTRe2). So, my opinion (and request) is unchanged. -- Tuválkin ✉ 20:45, 28 January 2012 (UTC)

It may be either   (vSTRe2) or   (vSTR2e), I have no preference and I’m sure some of you can decide better than I can which is better. -- Tuválkin ✉ 20:52, 28 January 2012 (UTC)

(And anyway, we all agree thst this is not a k-type Coumpound Junction icon, yes?) -- Tuválkin ✉ 20:45, 28 January 2012 (UTC)

In fact, given   (uvSTRlf) (we'll ignore   (vSTRlf) for the time being), then shouldn't   (vk-STR2) become   (v-STRl2)? Useddenim (talk) 19:11, 28 January 2012 (UTC)

I don’t know what the "l" is doing in that name: Left what? The relevant properties are left track of double line, shown by the placement of the root in relation to the dash "-", the 2nd corner, shown by the "2", and the basic topology, a straight line, shown by "STR". Should be simple, at least this one. -- Tuválkin ✉ 20:45, 28 January 2012 (UTC)

Ah! Meanwhile I “found”   (vÜWol) — whose normalized name is  (vSTR2), right? And therefore its left-only variant should be  (v-STR2), or am I completely wrong? -- Tuválkin ✉ 21:06, 28 January 2012 (UTC)

As an interim measure, I moved   (vk-STR2) to   (v-ÜWol), which is likely not the best name, but at least will not conflict with anything else. Perhaps Circeus has some thought on an alternate name. Useddenim (talk) 21:48, 28 January 2012 (UTC)

Sounds good. It is indeed the left line of   (vÜWol), however outdated that name may be, so   (v-ÜWol) is a good name for it. -- Tuválkin ✉ 06:30, 29 January 2012 (UTC)

Legend v-ÜWol uvSTRc1 ≠ v-ÜWol File:BSicon v-ÜWol.svg File:BSicon v-ÜWol.svg both icons overlaid

At en:Wikipedia:Route_diagram_template/Catalog_of_pictograms/parallel_straight_tracks#1.2F2_shift there’s a family of almost identical shapes:  (uvSTRc1), and its other 3 corner counterparts, and the 4 light blue versions. This is supposed to match up with   (uvSTRal) et c., so they are not geometrically identical to the left line of   (vÜWol) — but at low sizes I was almost fooled. (And of course that’s not a good name, and, even worse,  (uvSTRc1) ≠  (vSTRc1)!) -- Tuválkin ✉ 08:40, 29 January 2012 (UTC)

There are some DEEP problems in the column shift set. I have to get to those, but I kinda completely lost track of my naming proposals. Circeus (talk) 22:57, 31 January 2012 (UTC)

Given the nearly-identical appearance of   (v-ÜWol) and  (uvSTRc1) (colour aside), should the two groups be harmonized so they have the same geometry and match to each other? Useddenim (talk) 00:00, 1 February 2012 (UTC)

The geometry is similar, but the topology is completely different:   (v-ÜWol) results from (half of) a double line converging to corner 2 from center-top, while  (uvSTRc1) results from (half of) a double line going from center-top to center-bottom while shifting rightwards. If the 5:1 ratio of icon side to line width were different, this similarity would diappear. Besides, in practice, what you propose would not be easy to do right (either one or both icon families would need to be severly misshapen), even if it were theoretically sound. -- Tuválkin ✉ 09:26, 1 February 2012 (UTC)

v! icons

Moved from User talk:Tuvalkin.

Moved from Talk:BSicon/Renaming#v! icons.

So, I just spotted new icons with odd names from you. At first I thought this was a good place to reuse the b- prefix, then I realize these are BS2, not v- icons. Here's some replacement names:

•   (v!STRl2r) → BS2l2r (replacing vSTR of   with BS2)  (STR3h+1h)
•   (v!STRr2l) → BS2r2l  (STR2h+4h)
•   (v!KRZl2ru) → KRZo+BS2l2r or KRZ3+1oBS2  (KRZ3h+1ho)

Updated. Useddenim (talk) 18:41, 2 February 2014 (UTC)

I didn't quite get to elaborate my + system for overlapping roots, but it's already used with +BS platform icons. In any case you have the wrong suffix: the main feature is the vertical straight road, and it's going over the other one. If you don't think the BS2 connection is needed, then KRZl2ro will probably do the trick fine (v- implies it to be some weird w- style icon with a vSTRl2r crossing a vertical STR line.).

Alternatively one could use the corner numbers BS23+1, but that gets odd with a root that already contains a number at its end. Circeus (talk) 20:17, 5 October 2012 (UTC)

v!STRl2r v!STRl2r KRW+l¦KRWr vs. BS2r¦BS2+l

Actually, they are KRW icons offset by half a column. BS2 icons only shift a line over by half a column. Useddenim (talk) 01:34, 8 October 2012 (UTC)

KRWs and BS2s are the same basic thing, and each of them is half a SPL. I needed the icon, but all this made me hesitate about the icon names — I used an "!" to make it clear that a renaming is sought. So, I have no alternative ideas, not even specific suggestions about this one, only that names should be homogenous and intuitive (we all agree on that), and that replacing all BS2 with a less confusing name would be a great thing (but better to change it to something that is really good, not yet another kludge). Maybe we should move this discussion to a talk page, right? -- Tuválkin ✉ 02:26, 8 October 2012 (UTC)

Sounds good to me, though I'm not sure how much attention that'll get. I mean, we are the main guys interested in the names of these things. Circeus (talk) 15:26, 8 October 2012 (UTC)

Consider:

¼ shift	½ shift	¾ shift	full shift
(v-BS2l)	(BS2l)	(vBS2l-)	(v!STRr2l)
possible alternate roots
BS4	BS2	BS3	BS1

Useddenim (talk) 16:57, 8 October 2012 (UTC)

[undent] My proposal is to start by figuring out the name of the corresponding icon with two lines and work down from there with the dash system (although I dislike the dash system, it's probably not going away). If we start by not touching the BS2 icons (i.e. because of  (BS2lr) and its ilk, which would have to become weird variants on SPL), then vBS2l has to be  (BS2l※ BS2+r) ! Such an icon doesn't exist as far as I know. We can either keep the original single-line BS2 icons, create redirects from the new names, or move them with leftover redirects.

  (uvSTRal) (and the -l2r icons) become   (vBS2vl), with -v being the suffix I intended for left-to-right shifts within a v- icon. Originally I had   (vSTRr2l) as vSTRvl, but I'd rather keep these sets united, so it becomes v-BS2vl. Having redirects from the -r2l/-l2r roots is as workable as keeping the -l2r names, though. This finally deals with the problem that we got   (uvSTRal), yet   (vSTRal)! (both these sets should be curved  (vSTRa). IIRC Tuvalkin has made some of these using the SPL root.) It has to be noted that these icons are shifted forms of the  (BS2l※ BS2+r)  combo (i.e. vBS2l and vBS2vl correspond roughly to   (vSTRal) and   (vSTRel)), hence why I made sure both use the BS2 root.

As noted above, vBS2 is preoccupied (technically these icons don't exist yet, but I am convinced it won't be long... probably thanks to Tuvalkin, too), so although the   (vSTRal) set I wanted to deal with by using vBS2, I find myself having to use vBS4:

vSTRal		vSTRel		vSTRar		vSTRer
These names are based on the original names of the  (vSTRa) halves. Making up a "proper" system (e.g. starting from the middle to corner BS2 names, with the right line being the main one) always end up not working out
vBS4l		vBS4+r		vBS4r		vBS4+l
(vBS2+r-)	(vBS2l)	(vBS2+r)	(v-BS2l)	(vBS2r)	(v-BS2+l)	(exvBS2r-)	(vBS2+l)
vBS4l-	v-BS4l	v-BS4+r	vBS4+r-	vBS4r-	v-BS4r	vBS4+l-	v-BS4+l
vSTRalr				vSTRelr
BS2+r-BS2+l				BS2r-BS2l

I'd rather not touch the BS3 icons (especial that  (exv-STR4h)^[renamed] abomination) that are unique to Tuvalkin's special needs, but here's the result:

vBS2-+r	vBS2l-	vBS2+l-	vBS2-r
vBS3l- v-STR+4h	vBS3+r- vSTR2h-	v-BS3r vSTR+1h-	v-BS3+l v-STR3h
(Renamed. Useddenim (talk) 13:24, 27 May 2013 (UTC))

I'm still of the opinion that BS2 is a straightforward enough root for the "shifted KRW" of Tuvalkin's new icons, with several possibilities for the scheme. Although it is particularly grokkable to us, BS23+1/BS22+4 are very unseemly because of the two numbers next to each others. Circeus (talk) 13:43, 9 October 2012 (UTC)

New root

What about SHI for EN:SHIft/DE:verScHIeben? which then gives

¼ shift	½ shift	¾ shift	full shift
(v-BS2l)	(BS2l)	(vBS2l-)	(v!STRr2l)
SHI4	SHI2	SHI3	SHI

and eliminates the somewhat-confusing BS. Useddenim (talk) 16:58, 31 December 2012 (UTC)

Note: implemented by YLSS as

1×¼ shift	2×¼ shift	3×¼ shift	4×¼ shift
(SHI1)	(SHI2)	(SHI3)	(SHI4)

Useddenim (talk) 18:41, 2 February 2014 (UTC)

More column shifts

Before anyone asks: Yes, I created these, as I needed them for a diagram:

•   (exBS4l)   (BS4l)
•   (exBS4r)   (BS4r)

but I don’t insist on these names — indeed I have no clue, at least nothing cogent to add to what was already said in this matter. It is part of the big SPL & halfKRW discussion, which seems to be going nowhere. -- Tuválkin ✉ 21:15, 25 February 2013 (UTC)

As proposed at Talk:BSicon/Renaming/SPL#v! icons, I had suggest BS4 (or SHI4) for a quarter shift and BS1 (or SHI) for a full-width shift. Also,   (v!STRr2l) and   (v!STRl2r) already exist. (Watch out for that {{Duplicate}} tag!) Useddenim (talk) 00:45, 26 February 2013 (UTC)

Yep, and those two were made by me, even! As for SHI, well, why not, but then we’d need to aknowledge that KRWs are half SHIs, and that File:BSicon BS4r.svgFile:BSicon BS4l.svg (“SHIlr”) is the same basic thing as a  (vÜWB) stratched almost past the icon edges. Nothing wrong with that — after all, an ABZ is the same as two overlapped STRs with one common end… -- Tuválkin ✉ 04:51, 26 February 2013 (UTC)

But overlapped and stretched icons are v. different beasts! It's easy to come up with a system for naming icons that overlap, not so for stretched ones. Circeus (talk) 18:55, 6 March 2013 (UTC)

Yes, you’re right. The whole SHI / KRW / ÜWB / BS2 / SPL thing needs to be addressed holistically. I for one, have no bright ideas, I’ll go with any clever stuff you guys agree on. -- Tuválkin ✉ 22:20, 6 March 2013 (UTC)

.../parallel railways/uw/double

moved from Talk:BSicon/Categorization#.../parallel railways/uw/double

Also,   (v-STR2(!)) seems to be a duplicate of   (vÜWol-). YLSS (talk) 19:11, 13 June 2013 (UTC)

Yes, I didn’t notice   (vÜWol-) in time. I hope   (v-STR2(!)) will force the renaming of  (v-STR2) (to   (v-SPL2e) or some such). -- Tuválkin ✉ 12:51, 14 June 2013 (UTC)

I won't support that one, as long as we have the overall direction-of-travel philosophy. And I certainly don't see any (half of) split in it. I would rather say that   (vÜWol-) should be e.g.  (vSTR2-R): like  (VIADUKT-R) is half of  (VIADUKT), so   (vSTR2-R) is half of   (vSTR2). Although I must admit that -L/-R suffixes are getting overused (my fault here as well). (talk) 15:48, 14 June 2013 (UTC)

I think you’re utterly wrong there, and I’m very surprised.

• So, you cannot see how this is half of this  ?
• And do you defend that while  (vSTR) vs.  (vSTR-) is well named, however   (vSTR2) vs.   (vSTR2-) is not? Or do you think that  (vSTR-) should be renamed   (vSTR-R)?! (No, I don’t think that the analogy with viaduct or station halves is productive, but rather very misleading, needless, and against established practice.)

-- Tuválkin ✉ 21:16, 14 June 2013 (UTC)

Partly my fault, too. I never considered the possible confusion when I named  (v-STR2) (and I don't think that  (vSTR2) existed yet). Distasteful as it may be, we may have to consider resurrecting “ÜW”: i.e.  (vÜW2) etc.) Useddenim (talk) 01:56, 15 June 2013 (UTC)

No we don’t. You (plural) just could accept that SPLa/e (or STRa/e doesnt matter here) is an acceptable name for this kind of icons. But all this was discussed elsewhere. Lets keep it there, not in the page about categorization. -- Tuválkin ✉ 09:06, 15 June 2013 (UTC)

• Certainly, I can't! It's only after I read your suggestion that I can imagine it to be a part of split, but that's completely against WYSIWIG! Indeed that's not far away from regarding  (v-STR) as a part of   (ev-STRe).
• Heh, we seem to be paying attention to different aspects of the established practice. My way of interpreting things is like that: in  (vSTR-) I don't see a half of a  (vSTR), but rather two lacunae for parallel lines, occupied respectively by a straight track + nothing. In  (v-STR2) I see "nothing + track to the 2nd corner". In   (vÜWol-) I see "track to (? some oblique point) + nothing"; and that obliqueness can't really be explained in any other way than by imagining that it's half of a   (vSTR2) (yes, here we indeed have to imagine, no other way - that is neither "f" point, nor lower-right "k" point, nor right parallel track point).
• I'm both hands up for SPL for icons that do show two lines converging or diverging. It can be done without introducing a new root (e.g.  (vSTRf-STRf)), but I certainly prefer SPL to STRsmth for such icons.
• I'm also against ÜW.YLSS (talk) 13:01, 15 June 2013 (UTC)
• And most surprisingly, you used the "-L" modifier yourself in  (hvSTR-L+1-), less than a month ago! Also, the series  (hv-STR+1) +  (hvSTR-L+1-) \  (hv-STR3) +  (hvSTR3-) presently occupies the names that should be reserved for elevated versions of either  (v-STR3) or separated   (vÜWor). Curving by 45° of two parallel tracks that run in different cells presents yet another geometry, incompatible with the   (vÜWor) series; and I have no idea whatsoever how to name them... YLSS (talk) 20:52, 15 June 2013 (UTC)

BTW, shouldn't   (v-STR2(!)) and   (vSTR2-(!)) be switched? YLSS (talk) 15:48, 14 June 2013 (UTC)

Yes, that’s a mistake I didn’t fix after I noticed it because these two are going to be renamed, anyway. -- Tuválkin ✉ 21:16, 14 June 2013 (UTC)1

The problem: two dissimilar geometries that can be vaguely described by the same name.
The solutions: different but similar names. p is currently unused, and is certainly the obvious and multilingual prefix for “parallel”.

(vSPL2)   (vSPL2)	(vSTR2-)	(v-STR2)
(pSTR2)	(pSTR2-)	(p-STR2)

Useddenim (talk) 11:48, 16 June 2013 (UTC)

Heh, smart ;) Maybe rather "w" or double-vee: "vv" ? BTW, since "SPL" root implies that on one side there are parallel lines (but on the other there aren't), isn't "v" excessive? That is,   (SPL2). YLSS (talk) 12:30, 16 June 2013 (UTC)

Yeah, no "v"; but let's just stick with the mnemonic, as we already have the prefix W for water. Useddenim (talk) 16:36, 16 June 2013 (UTC)

I dont have time to reply right now to this recent installment of the debate (although it is just repeating of what I said above, long ago), but let me remind you that "p" is already taken, as in  (uxpBHF),  (upBHF), and  (uexpBHF). Either way, it seems a bad way to deal with the “problem” (which for me is no problem at all, really), as we’d never know if any icon showing double lines should be  (vSTR) or “ (pSTR)” -- Tuválkin ✉ 19:28, 16 June 2013 (UTC)

---

Maybe not a very good idea, but still an idea... Present   (vÜWol) & co. are designed specifically for parallel tracks that concurrently shift by a full cell to the left over a span of two icons in height. They do not approach corners at exact 45°, and   (vÜWc1) etc. are not straight lines but sophisticated curves. Evidently, they would be unusable together with a possible parallel-line version of  (STRl+4). So, unless anybody wishes to revise their geometry, maybe it would be appropriate to name them   (vSHI4x2l),   (vSHI4x2c1),   (vSHI4x2+r),   (vSHI4x2l-) etc.? "x" here means "four quarters by two cells", and could possibly be replaced by another symbol ("×" ?). (BTW, a curiosity:   (v-STR2a) looks pretty much the same as  (SHI3l).) YLSS (talk) 00:49, 23 November 2013 (UTC)

Conundrum

moved from Talk:BSicon/New icons and icon requests#Conundrum

I am at a total loss as to how to name these:

• 5001: Upper parallel line to left parallel corner 4
  5001: Upper parallel line to left parallel corner 4
• 5002: Left parallel line to right parallel corner 2
  5002: Left parallel line to right parallel corner 2
• 5003: Lower parallel line to right parallel corner 2
  5003: Lower parallel line to right parallel corner 2
• 5004: kompound junction corner 2 to right parallel line corner 2
  5004: kompound junction corner 2 to right parallel line corner 2

They have elements of both SHI2 as they cross between left and right parallel lines, and VSTRc# corners. See en:Template:Podanur–Pollachi branch line for usage. Useddenim (talk) 17:04, 4 May 2014 (UTC)

Hmmm... I wanted to discuss this question for a while, so I guess now is indeed the time... Also, I would like everybody to comment at Talk:BSicon/Icon geometry and SVG code neatness#vÜWor vs. vSTR3 (near the end) before uploading any new icons.

First, I'd like to note that   (vSTR3+g-) is a duplicate of  (vSTR3-R) (with a bezier curve instead of a circle arc), and I hope you don't mind that I renamed v-STR3+l to  (vSTR3+l-L) basing on  (uvSTR3+l). (v-STR3+l should IMHO show something like

.)

Such naming pattern was first proposed by me in the section above. The idea is like that: for icons that are not quite meaningful on their own, but only as a half of a pair of parallel lines, we can employ -L & -R suffixes. Thus,  (vSTR3) =  (vSTR3-L) +  (vSTR3-R), in contrast to  (vSTR3-) &  (v-STR3). (Tuválkin, I know that they may be named using SPL, but that is too clumsy for such simple icons.)

WRT present icons, I would suggest the following:   (5002) →   (vSTR3+R),   (5001) →   (vSTR3+l+L),   (5003) →   (vSTRr+1+R). IIRC somebody proposed somewhere to use +L & +R for icons like this:  (leer+hl), and I think it can be extended to the present case as well... Cf.  (hSTR-R)  vs.   (hvSTR+R) with  (vSTR3-R) vs.   (vSTR3+R).

WRT   (5004): OMG... better delete this one... While the right half is the same as in   (5003), the left one is a cross between the lower slot for parallel lines and a k-curve. I don't endorse such icons, because the whole purpose of k-icons is a gentle curve, as nearly circular as possible, so it would be pointless to connect it to anything else than  (kSTRq+l). In most cases, it is possible to use a corresponding ÜW-curve and everything becomes simpler. I did this in en:Template:Podanur–Pollachi branch line, and I hope that the reduced space between the elevated line and the ground-level lines won't be a problem?..

I have also added a note at #vKRWs below.

-- YLSS (talk) 19:46, 4 May 2014 (UTC)

Now that I moved parallel diagonal lines a bit closer, these icons are in need of redrawing. However, there is even less space for the curve now, and don't see how all that can be tolerably drawn. YLSS (talk) 14:27, 8 June 2014 (UTC)

Parallel lines ¼ shift and ½ shift

moved from Talk:BSicon/Icon geometry and SVG code neatness

Compare   (vSTRel) vs.   (uvSTRel) et al. Useddenim (talk) 22:15, 30 July 2013 (UTC)

That’s not a matter of geometry, but one of naming. -- Tuválkin ✉ 01:20, 31 July 2013 (UTC)

Now replaced by  (vSHI1l) &  (uvSHI2l) etc. 128.205.46.134 18:44, 4 December 2013 (UTC)

At last, yay! We can archive this now… -- Tuválkin ✉ 20:47, 4 December 2013 (UTC)

Heh... Don't forget about the essential  (vSTRc1) &  (uvSTRc1) ! YLSS (talk) 22:24, 4 December 2013 (UTC)

SPL vs. SHI

As you may know, I've started renaming various shifts to the "SHI" root, and as yet faced no criticism. So now I'm gonna ask: do we need the SPL root? We have other options now:

•  (vSTRa) →  (vSPLa) or   (SHI1lr) (aka   (vSHI1+l-SHI1+r))
•  (vSTRe) →  (vSPLe) or   (SHI1+lr) (aka   (vSHI1l-SHI1r))
•   (vSTRe-) →  (vSPLe-) or  (vSHI2g+l-)
•  (vSTRl) →  (vSPLaq) or   (SHI1lrq) (aka   (SHI1+rq-SHI1+lq))

In case of splits on curves, SPL is certainly shorter:

•   (vABZaq+r) →   (vSPL+ra) or   (vSTR+r-STR+r) or   (vSTR+r-ABZq+r)
•   (STR+r-ABZg+r) →   (vSPL+re) or   (STR+r-STR+r) or leave as it is
•   (vSTRl+4e) →   (vSPLl+4e) or a name that would certainly present more problems

As Circeus wrote above, SPL root "hardly addresses the problems that STRa/e already has". Indeed, it employs the "a"/"e" suffixes that potentially may be needed for other purposes (and employing "f"/"g" instead won't solve the issue). However, SPL shortens names. So I'm quite neutral on this topic. What do others think? YLSS (talk) 09:07, 27 November 2013 (UTC)

I vote for SPL (but then I was the one who originally proposed the name), as I think it’s clearer than a compound SHI. However, is the “v” prefix necessary? Also, as an alternative to the a/e suffixes, what about the unused letters s (en/de: start) and z (de: zipfel)? Useddenim (talk) 19:09, 27 November 2013 (UTC)

Nah, more complex names are certainly better than new letters. But the issue of "a"/"e" is not critical, no one has created such an insane icon yet that would require different meanings of start/end, so I would be pretty happy with SPL. YLSS (talk) 22:54, 27 November 2013 (UTC)

I vote for SPL in all cases where it yields shorter filenames. Look, surely there’s common traits shared by SPL and SHI, but the same can be said about ABZ and STR, and yet we do not consider getting rid of ABZ just because  (ABZrg) can be unambiguously labelled as two STRs and a "+" based on that it is the same as  (STR※ STRrg) . -- Tuválkin ✉ 04:31, 7 December 2013 (UTC)

OK, OK, so it's SPL then. (Actually, I think I wrote a couple of times that "ABZg+l" is redundant to "STRg+l", but of course that is a purely theoretical discourse.) YLSS (talk) 08:34, 7 December 2013 (UTC)

Do we agree that SPL takes precedence over SHI1 in the naming of single/parallel transition icons (e.g.   (utvSPLe-SHIl) etc.)? Useddenim (talk) 13:15, 8 December 2013 (UTC)

You mean like   (vSPLa-SHI2l)? Well, yeah, that's why I named it like that. YLSS (talk) 16:50, 8 December 2013 (UTC)

I actually meant   (utvSPLe-STRl). Useddenim (talk) 17:32, 8 December 2013 (UTC)

Which, upon reflection, should probably be  (utvSHI1l-STRl). Useddenim (talk) 17:36, 8 December 2013 (UTC)

It certainly should. BTW, now I see that I was right in creating an army of redirects: if we take away the second half of   (utvSHI1l-STRl), we do get  (utvSHI1l-), even though the file is located at  (utSHI1+r). YLSS (talk) 18:02, 8 December 2013 (UTC)

Hm... But that leaves the question whether we should move e.g. (  (vSTRlf) →)  (vSTRgl) →  (vSHI1l-STRl). So with that shortcut we are not happy? YLSS (talk) 18:10, 8 December 2013 (UTC)

IMHO   (vSTRlf) should be a redirect to   (vSTRl), and  (vSTRl) be released in favour of   (SPLl); but you are correct in suggesting that  (vSTRgl) be renamed  (vSHI1l-STRl). Useddenim (talk) 02:10, 10 December 2013 (UTC)

WRT "vSTRlf", I've started a topic at the main talk; "vSTRl" is already  (SPLaq); and I do not suggest renaming  (vSTRgl), only that if it's going to be renamed, then to  (vSHI1l-STRl). But does it need to be renamed? I'm against introducing new "shortcuts", but I can live with old high-use names unless they are misleading or conflicting or ugly. YLSS (talk) 06:48, 10 December 2013 (UTC)

SPL vs. vSPL

I’m now convinced that a simple “fork” should be SPL, not vSPL. Here’s why:  (vSPLa). -- Tuválkin ✉ 04:31, 7 December 2013 (UTC)

Yup, I came to the same conclusion some time ago, cf.  (SPLaq). A good icon ↑, BTW, I recollect a couple of cases where it would've been handy. YLSS (talk) 08:34, 7 December 2013 (UTC)

I didn’t meant to say that nobody else noticed this solution to the SPL vs. vSPL issue, indeed it was the  (SPLaq) in your many colors page which gave me the idea about the double fork. -- Tuválkin ✉ 15:41, 7 December 2013 (UTC)

Also, I assume that in order not to create "SPL1", "SPL2" etc., we can postulate that this root is only used for parallel-to-single and single-to-parallel cases, and not for such things as  (BS2lr). YLSS (talk) 08:34, 7 December 2013 (UTC)

 Agree. Useddenim (talk) 14:12, 7 December 2013 (UTC)

 Me too. While things like  (BS2lr) need to be renamed, they surely are not single⇄double forks as SPL are. -- Tuválkin ✉ 15:41, 7 December 2013 (UTC)

  (SHI2lr)? (I think   (SHI22h3h) or   (ABZ2h3h) would be going overboard.) Useddenim (talk) 17:13, 7 December 2013 (UTC)

I think we should use the "h"-suffix only when there are no other ways to name an icon, as in  (STR2h+r) (not a shift) or  (STR2h+4h) ("SHI2+4" "SHI42+4" is possible, but messy), I have even renamed the series of   (ABZg2h) icons that I had introduced previously to  (SHI2gl). "things like  (BS2lr) need to be renamed": well... go ahead, if you want. I intended only to rename parallel-line shifts, which show different geometry, and to leave BS2s and KRWs as hysterical raisins. YLSS (talk) 19:06, 7 December 2013 (UTC)

(Comment) I like the way this page currently looks like! YLSS (talk) 12:03, 16 December 2013 (UTC)

vSPLa-

Hmm... I intended to move   (vSTRe-) to  (vSPLe-), especially since I uploaded  (SPLeq-) some time ago, but wait! We now have  (vSPLe), and if we take the left (IRL right) half away, we get something like  . WTF? YLSS (talk) 11:16, 9 December 2013 (UTC)

Since the e/x variations can’t properly cope with the four elements of this icon, I’d say leave well enough alone. Useddenim (talk) 02:10, 10 December 2013 (UTC)

Erm... what? where? when? I'm afraid I do not understand at all what you mean. That is, pertaining to   (vSTRe-) →  (vSPLe-) (or  (vSHI2g+l-)). YLSS (talk) 06:24, 10 December 2013 (UTC)

Having thought even more, I assured myself that pure logic and symbol arithmetic necessitates  (vSPLa) =  (vSPLa-) +  (v-SPLa), and since I found at least one diagram where they were handy, I uploaded them. So   (vSTRa-), which is moreover not a single-to-parallel transition in the sense defended by Tuválkin above, should be named somehow otherwise. Any opposition to  (vSHI2gl-)?  (SPLeq-) should then become   (SHI2q+r-), I think. YLSS (talk) 13:14, 10 December 2013 (UTC)

1. But equally true,  (vSPLa) ≡  (vSHI1r-) +  (SHI1+r) +  (SHI1+l) +  (v-SHI1l), which are the four elements I meant.
2. Applying the KISS principle, perhaps   (vSTRa-) →‎   (vÜSTl-) instead of  (vSHI2gl-) (cf.  (vÜSTl)—one suffix character vs four? Useddenim (talk) 12:57, 12 December 2013 (UTC)

I arrive late to this party, but seems that seeing this as an incomplete ÜST is the best approach. I feel guilty I created all those vSTRa and vSTRe back then, mostly unused, then lobbyed to have them renamed as SPL, and then launched the  (vSPLa) problem. However, seems that both SHI and h made the naming system become solid enough to be able to cope with all kinds of unexpected stuff — even though in this case, I think,   (vSTRa-) →‎   (vÜSTl-) is the best idea. -- Tuválkin ✉ 14:12, 12 December 2013 (UTC)

(I thought you guys proposed in older sections above that SHI would be the ultimate replacement for STR/BS2/KRW/ÜST/ÜWB? So what is it now, cold feet?) Personally, I'm totally against multiplying instances of a root with Ü in it.  (SHI2gl),   (vSHI2gl-),  (v-SHI2gl),    (vSHI3gl-) etc. etc. (Tuválkin, nothing to feel guilty about! Very useful icons, and if you now regret how you named them, well, we all have renamed our own uploads, possibly me foremost. If any sense of guilt remains, you can help with renaming ;) ). YLSS (talk) 16:42, 12 December 2013 (UTC)

4x4 vSPLa

Maybe like this?

• ··vSPLa···· :  
• ··vSPLaxl·· :  
• ·xvSPLa···· :  
• ·xvSPLaxl·· :  
• e·vSPLa···· :  
• e·vSPLaxl·· :  
• exvSPLa···· :  
• exvSPLaxl·· :  
• ··vSPLa··xr :  
• ··vSPLaxlxr :  
• ·xvSPLa··xr :  
• ·xvSPLaxlxr :  
• e·vSPLa··xr :  
• e·vSPLaxlxr :  
• exvSPLa··xr :  
• exvSPLaxlxr :  

If so, then  (exvSPLa) and  (exvSPLe) should be renamed. -- Tuválkin ✉ 14:54, 12 December 2013 (UTC)

I don't like the idea of "l" and "r" appearing unexpectedly with "x" suffix when there were none without it. IMHO,   (vexSPLa-xSPLa) is far more understandable (and re-coinable) than   (evSPLaxlxr). But: is the game worth the candle? YLSS (talk) 16:21, 12 December 2013 (UTC)

We're just asking for trouble if we go down that (rail)road. IMHO, those icons would be used so infrequently they should just be created with overlays on an ad hoc basis. 128.205.48.116 16:28, 16 December 2013 (UTC)

Sure, I don’t seriously mean to upload these — not now, at any rate. The above proposal list was just meant to address Useddenim’s concern above (at 02:10,2013.12.10), showing that our current system can name all usedness variations for all four segments in these icons. (Not so sure about "m" versions, but we really don’t need to go there.) -- Tuválkin ✉ 20:52, 16 December 2013 (UTC)

vBHFa

Any desire to rename  (vBHFa) etc.? If so, any thoughts on a better name?   (SPLa+BHF) (SPLa+vBHF) (ROOT + ROOT)? YLSS (talk) 19:26, 7 December 2013 (UTC)

I’d prefer  (SPLa+vBHF), it assumes less. (After all,   (SPLa+BHF) could be something like  …) And, of course, the name  (vBHFa) should be something like   (vBHFa(!))… -- Tuválkin ✉ 20:50, 7 December 2013 (UTC)

Of course with "v", that was a slip of attention. (Aargh! Delete that icon before I start seeing in my nightmares! ;) It is do-able in the "u" set:  . ) YLSS (talk) 21:57, 7 December 2013 (UTC)

Split+junction

Should we rename icons showing a split with a junction, like  (vSTRag+l)?  (SPLa+l) is for a split on a curve, so I can only suggest   (SPLa+v-STR+l). But that is long; the current name, on the other hand, is relatively explicable. Names for the other subclass, e.g.   (vSTRaq+l), contain an enigmatic "q", and I would favour changing them to   (SPLa+l+SHI1r). Opinions? Any other ideas? YLSS (talk) 10:09, 13 December 2013 (UTC)

 (SPLa+l+g)? Useddenim (talk) 13:35, 16 December 2013 (UTC)

Why not. YLSS (talk) 13:43, 16 December 2013 (UTC)

I don’t understand that "+g": «From gegen»? What would that be? (Maybe  (WSLa)…?) Shouldn’t it be "+f" instead? -- Tuválkin ✉ 18:30, 16 December 2013 (UTC)

See the vWSL set. 128.205.48.20 18:44, 16 December 2013 (UTC)

Thanks for the tip. Mighty useful, as I created and named those icons myself. -- Tuválkin ✉ 20:22, 16 December 2013 (UTC)

And various ÜW-junctions. "f" should be reserved for the center of the bottom line. YLSS (talk) 19:20, 16 December 2013 (UTC)

I get it: As a location, not a direction, "g" is the midpoint of the top edge of the icon and "f" is the midpoint of the bottom edge of the icon. Okay, that makes sense and is concordant with all the rest. -- Tuválkin ✉ 20:22, 16 December 2013 (UTC)

vKRWs

moved from User talk:Useddenim#vKRWs

moved to Talk:BSicon/Icon geometry and SVG code neatness/Archive 3#vKRWs

Crossings

@Useddenim and Tuvalkin:

• Should  (vKRZoa)/e be KRZqSPLau (like  (KRZq+1u) etc.), or something else?
• Based on that, what would  (uvSPLe+G4qo) (etc.),  (mKRZuSHI1lq) be? Would we need two qs for the latter?

Jc86035 (talk) Use {{re|Jc86035}}
to reply to me 07:52, 19 March 2017 (UTC)

• I think that  (vKRZoa) etc. (for whose bad naming I wish to apologize in retrospect), could be more sensibly renamed SPLaoKRZ or even SPLa+STRuq, while for me  (mKRZuSHI1lq) should be something like  (mSHI1loKRZq) — that is, with the trailing q acting as a final quarter-turn rotation switch for  (mSHI1loKRZ)  (but with the right bridge): When two qs may be needed we know there’s a problem with the name, as they would sort of cancel each other. -- Tuválkin ✉ ✇ 10:53, 19 March 2017 (UTC)
  • Allowing o/uKRZ(q) to be tacked on to basically anything is a good solution, although I'm not sure where the suffixes would go or how we'd handle junctions (like  (KRZol)). Jc86035 (talk) Use {{re|Jc86035}}
    to reply to me 11:18, 19 March 2017 (UTC)
  • Would we handle  (vSTRr-KRZ) the same way (vSHI1lKRZ-STRr)? Maybe it would be better if u/o were placed after KRZ instead of before. Jc86035 (talk) Use {{re|Jc86035}}
    to reply to me 11:22, 19 March 2017 (UTC)

You could use the KRZ ÜW family ( (KRZq2+4u),  (KRZr+1u), etc.) as a guide for naming. Useddenim (talk) 11:44, 19 March 2017 (UTC)

@Useddenim: The problem is that with  (mKRZuSHI1lq) the SHI1lq part needs to be before KRZ to avoid ambiguity as to which part of the icon q refers to. Perhaps it could be before KRZ with something crossing a vertical line and after KRZ with something crossing a horizontal line? Jc86035 (talk) Use {{re|Jc86035}}
to reply to me 14:00, 19 March 2017 (UTC)

That’s only a problem if you chose to ignore what I’ve been saying for years now: Let «the trailing q act»« as a final quarter-turn rotation switch». No ambiguity if it always applies to the whole icon. -- Tuválkin ✉ ✇ 12:47, 10 April 2017 (UTC)

@Tuvalkin: I suppose we could do that, but it would probably have to exclude parallel lines and icons with the plus sign, like  (STR+GRZq), for the convenience and sanity of everyone else making RDTs. Jc86035 (talk) Use {{re|Jc86035}}
to reply to me 13:58, 10 April 2017 (UTC)

a/e suffixes

@Useddenim, Tuvalkin, and Lost on Belmont: Are the a/e suffixes necessary for 45° splits ?  (SPLe2+4),  (SPLa3) etc. could be renamed SPL4+2, SPL3[+g] (like  (CONT4+2),  (CONT3+g)). (Note that  (SPL2+1) already exists.) Would +g/f be necessary?

You make a good point, since there is an explicit direction with these icons. So ‘No’. Useddenim (talk) 10:23, 10 April 2017 (UTC)

In addition, shouldn't  (ex-SPL2+1) be exSPL2+1-L (since there's no v prefix)? Jc86035 (talk) Use {{re|Jc86035}}
to reply to me 09:45, 10 April 2017 (UTC)

With the SPL icons there is an implied v, so I would say stick with the simpler, more common, existing name (unless you can come up with a conflicting example where the ‘no v but -’ construct for half parallel across icons would create confusion. Useddenim (talk) 10:23, 10 April 2017 (UTC)

@Useddenim: The problem is that there's also  (vSPLa)/e, so if we were to do that sort of thing then vSPL[a]2+4-R and vSPL[a]2+4- would be different. (Would vvSTR2+4 be three parallel tracks?) Jc86035 (talk) Use {{re|Jc86035}}
to reply to me 10:35, 10 April 2017 (UTC)

@Jc86035: Whaft would a diagonal vSPL connect to? Useddenim (talk) 18:45, 10 April 2017 (UTC)

@Useddenim: Probably the parallel corner tracks, since all diagonal tracks are equally spaced ( ). Alternately could connect to   (vSTR-LR). Jc86035 (talk) Use {{re|Jc86035}}
to reply to me 04:07, 11 April 2017 (UTC)

  (5001) = vSPLl+3-R;   (5002) = vSPL2+g-L;   (5003) = vSPL1+r-L? (This doesn't cause any conflicts for  (ex-SPL2+1), but I still think it shouldn't use the parallel lines syntax because using the dash without v you would expect something in the bottom half of the icon, cf  (v-BRIDGE) and  (-BRIDGE).) Jc86035 (talk) Use {{re|Jc86035}}
to reply to me 05:02, 11 April 2017 (UTC)

@Useddenim:  Resolved by only matching the inner curve and not the outer curve. Files renamed. Jc86035 (talk) 08:47, 14 August 2017 (UTC)