Talk:National Register of Historic Places listings in Hartford County, Connecticut

Time for a split.
The article is now 100k, and very sluggish when editing. The last time it got too large, the solution was to break out the City of Hartford as a separate page. This time, the logical split isn't that obvious, as there is not single location that has a meaningful percentage of the total page.

Some of the larger towns/cities (in term of entries):
 * Southington 38
 * West Hartford 32
 * Windsor 41

No one of these will reduce the article enough to make it worth the effort.

I considered some regional breakout, e.g. Windsor/Windsor Locks/Suffield, but I don't really like the idea. I'm leaning toward thinking we should identify the top four or five towns, and break out an article for each one. However, I'm hoping someone with experience in these matters can weigh in.-- SPhilbrick  T  14:41, 21 September 2010 (UTC)
 * I think splitting off a town is what's generally been done for these county lists when they're too long. Making separate lists for each of the three towns you listed looks like a good idea. --Polaron | Talk 14:47, 21 September 2010 (UTC)
 * Support Splitting Windsor. It makes the most sense as it has the most entries and would reduce the length. I would then propose reducing some of the images... for example, Trinity Methodist Episcopal Church is over 8MB, which is IMO more likely why the page is sluggish! Markvs88 (talk) 14:52, 21 September 2010 (UTC)


 * (ec) I've thought that about 30 to 100 is a rule of thumb of how many to have, for one of these list-articles, when pics and descriptions with sources are being added. Often splitting out towns is what has been done, e.g. just recently to split out Greenwich and Stamford out of Fairfield County list, in addition to Bridgeport which had been split.  But that approach could make the remainder more and more of an odd assortment.  It could be better to impose a partition, based on some meaningful groupings of towns or geographic divisions by rivers or highways.


 * For various NRHP list-articles splits have been done by official regions (e.g. Philadelphia or by divisions on northeast vs. south etc. via more arbitrary dividing lines (e.g. Denver and for Manhattan which was split by major streets 14th, 57th, 125th i think. There's at least one Pennsylvania county split into geo sections that way.  Would it make sense to group Windsor and Windsor Locks together, for example.  Would northern vs. southern vs. eastern vs. western Hartford County make 4 useful groups?  It takes looking at the linked google/bing map and imagining dividing that up into sections.  It's good you're asking now, it is time to do something here. --doncram (talk) 15:05, 21 September 2010 (UTC)
 * I would not support splitting out a couple of towns together as this doesn't pass any real litmus test -- would we split out Fairfield, Connecticut and Bridgeport, Connecticut together? I wouldn't, nor would I split Windsor, Connecticut and Windsor Locks, Connecticut. The thing about CT is that we haven't had county government in nearly 50 years, so they don't really mean much except for real estate purposes to most people. For months I've been making town categories like Category:Windsor Locks, Connecticut because the next entity above the town is the state. Anything else is arbitrary and not helpful to the population at large, IMO. Best, Markvs88 (talk) 16:02, 21 September 2010 (UTC)
 * I was thinking it might make sense to include Windsor Locks with Windsor, but you've convinced me, this is likely to create more confusion.-- SPhilbrick  T  18:44, 21 September 2010 (UTC)
 * Support breaking out Windsor, Southington, and West Hartford. I don't think doing Windsor alone is worth the effort. I support breaking out Windsor, Southington, and West Hartford. As Nyttend's list here suggests, those three are roughly similar, and the next largest is much smaller.-- SPhilbrick  T  15:38, 21 September 2010 (UTC)
 * Why wouldn't it be worth it? I do not understand. Best, Markvs88 (talk) 16:02, 21 September 2010 (UTC)
 * Because breaking out Windsor alone would reduce the page from 289 to 258, or down only 8%. The difference in size and editing speed would be barely noticeable, and we would be looking at a new split in a few months. I don't know exactly how to do the split, but I'll bet it is not much harder to break out three than to break out one, and that reduce the size enough to last for another couple years.


 * Plus, as soon as we do something other than the logical Hartford vs rest of county, we should clearly note on the page what is in what page, and once one has any single town broken out, one might as well have a list.-- SPhilbrick  T  16:17, 21 September 2010 (UTC)

Alternative proposal
I suggest removing all of the entries from the Hartford County page, and turning the Hartford county page into a summary page, that links to as many articles as are needed. That approach would reduce the size of this page materially, so it would load quickly. Then each reader would decide which of the individual pages they choose to open. At the moment, if you are interested in place in the City of Hartford, but don't know the structure, you might start with this page, which means you have to suffer through the complete load time of this page, only to realize that you don't care about a single entry on this page, and need to load another large page.

(I considered the extreme example of making it a dab page, but I don't think that is warranted, as this page is the logical place to discuss the aggregate counts of NRHP listing in the entire county, and to mention the NHL pages.)


 * Hartford County
 * Summary information on entire county
 * Pointers to National Historic Landmarks pages
 * Pointers to City of Hartford, individual towns (as selected), and rest of County
 * No table

Then I see individual pages as follows: -- SPhilbrick  T  17:01, 21 September 2010 (UTC)
 * Hartford (city)
 * Summary information on area covered
 * Table of entries
 * Southington
 * Summary information on area covered
 * Table of entries
 * West Hartford
 * Summary information on area covered
 * Table of entries
 * Windsor
 * Summary information on area covered
 * Table of entries
 * Rest of Hartford County
 * Summary information on area covered
 * Table of entries

a) Instead of "Rest of Hartford County" consider "Other locations in Hartford County". . b) split out all locations with 10+ sites. Bristol (14), East Hartford (10), Farmington (11), Glastonbury (11), New Britain (17), Simsbury (17), Suffield (11) = 91, or 32% of the 282 entries.
 * Support. I can support this, but would also make these suggestions:

Secondary alternative proposal
Consider leaving everything with < 10 sites not broken out. If we split as suggested above in suggestion b (91) + Windsor (41) + West Hartford (32) + Southington (38) = 202. That's 71% of the table! Thoughts? Markvs88 (talk) 19:40, 21 September 2010 (UTC)


 * I find that short lists are just a pain, where they have been oversplit in some other areas. It causes more work.  Most editors involved with Massachusetts have felt it was oversplit, probably still discussion showing from Talk:List of RHPs in MA.  One way it causes more work is you have to identify places that straddle borders, and list in both articles, and explain subtraction of duplicates in overall tally tables.


 * I personally would rather work on developing info in longer lists, in 30-100+ range, going up to 120 or 130 is okay. If you are visiting the area and going to identify places on a map, having them all in larger lists works:  you can see all of a given area in one linked Google/Bing map.  You cannot compare where places are, in one map, if small patches are taken out for 10 here, 10 there.  Also that is why i favor grouping the towns south of Hartford, vs. west and north of Hartford, vs. east of Hartford say.  Say the East of Hartford chunk would be defined as all parts east of the CT River, including: East Hartford (10), Manchester (4), Glastonbury (10), South Glastonbury (1), and others.


 * Another example perhaps comparable to Connecticut is National Register of Historic Places listings in Puerto Rico, where there are no counties, not even historically, the island is just divided into 70 or so municipalities. I didn't know much about it before working on the NRHP list-articles there.  The way it is now organized, with input of several editors there, is by regions that group the municipalities into chunks labelled northern, western, central, southern, eastern, + San Juan & metro area.  The division is not entirely arbitrary;  we found some unofficial tourism map, then later an official state tourism map, that divided them up into approximately equal chunks.  I wonder if some Connecticut state or Hartford region map with tourist divisions can be found, that would provide an equivalent partition into groups of towns here.  There's no rush, is there?  Couldn't we browse around and consider the options for a few days. --doncram (talk) 21:53, 21 September 2010 (UTC)


 * P.S. Depending on how towns are split out, some items that would appear in more than one would be: Glastonbury-Rocky Hill Ferry Historic District, Farmington Canal-New Haven and Northampton Canal, perhaps more historic district ones or bridges that span between towns. Maybe not too many.  Just to vent a small amount:  one further complication is that there's been contention, actually, for NRHP list-articles which include coverage of a district spanning into another NRHP list-article's area, about whether the first list-article's location info is allowed to mention the fact that the district extends into the other area.  (This came up at Talk:National Register of Historic Places listings in New Haven County, Connecticut, and recently at wt:NRHP where it was not really resolved before being archived.)  I hope editors here would be reasonable and allow for full description of an item in whichever lists it appears, but the more splitting you do the more of this overhead you have to deal with. --doncram (talk) 22:06, 21 September 2010 (UTC)
 * If 10 is too low, then maybe 15? But 30 seems a bit high for CT towns (and heck, the cities too). It's in one group now which is fine to me, but I cannot support any arbitrary groupings. Personally, I'd rather do away with the county groupings since they don't really exist anymore and have everything by town and state only. Markvs88 (talk) 22:41, 21 September 2010 (UTC)
 * At one level, I like doncram's approach, which could produce a small number of distinct pages of roughly equal size, but I don't think there's any natural allocation.


 * Compromising at 15 as a break would produce the following approximate (updated to reflect Simsbury) counts:
 * 127 Hartford
 * 41 Windsor
 * 39 Southington
 * 32 West Hartford
 * 17 New Britain
 * 16 Simsbury
 * 142 Other locations in Hartford County


 * I agree with Markvs88 that the county division is close to useless, but it did have the advantage of producing a manageable compromise between the state level and the town level.


 * I also agree with doncram that there's no rush. Doesn't need to be done this week or this month. If some other state has identified a better solution, let's steal it.-- SPhilbrick  T  23:33, 21 September 2010 (UTC)

Third alternate proposal
right|thumb|Map with towns Here's a map of the county by town. Perhaps group as west: West Hartford, Farmington, Burlington, Avon, Camden, Simsbury, Bloomfield. Those east of the river are Enfield, East Windsor, South Windsor, Marlborough, Manchester, East Hartford, Glastonbury). Remaining are in northern group (Windsor, Windsor Locks, Granby, East Granby, Suffield), and remainder is a southern group (Bristol, Plainfield, Newington, New Britain, Berlin, Rocky Hill, Wethersfield).    Does the Hartford Courant have news coverage / news delivery areas termed "northern suburbs", east, etc., organized roughly like this?  I would be happy for this to be modified to match up to any grouping like that.  By the way, I am noticing that the current list includes places that are smaller than towns, while some other CT NRHP lists town/city column was revised to strictly use town names. --doncram (talk) 22:18, 21 September 2010 (UTC)
 * Unfortunately, there is really no good way to group towns together. The only possible grouping that's somewhat well-known is the Central Connecticut area consisting of Southginton, Berlin, New Britain, Bristol, Plainville, Burlington. The rest of the Hartford area is pretty much monolithic. Also, this list in its current form does use towns as the location. There is no subtown entity listed in the current list. What specifically are you referring to when you say "places smaller than towns"? --Polaron | Talk 22:31, 21 September 2010 (UTC)
 * Oh, i was looking actually at the tallies by Nyttend by town at linked page, where there are non-towns like "South Glastonbury: 1, Unionville: 1, Warehouse Point: 1, Hartland: 2, and so on. I thot he was working from the current list-article.  But, right, Polaron a while back did change this Hartford County list-article to give just towns, not smaller locations that were originally in the list-article.  So, then, the tallies for towns that we were relying upon in this discussion, will be higher than counted by Nyttend. --doncram (talk) 22:57, 21 September 2010 (UTC)
 * Oppose. I just cannot support this as it is arbitrary. Markvs88 (talk) 22:35, 21 September 2010 (UTC)
 * I am not bothered by an objective though arbitrary division, but i understand this type of division doesn't appeal to everyone. The oversplitting causes other problems, IMO.  I don't care much though;  if a few editors here want to go with the by-town approach that is fine by me.  I am just trying to share some from my being aware how splits were done elsewhere. --doncram (talk) 22:57, 21 September 2010 (UTC)

Fourth alternate...........
Arrgh...wrote this and lost it once.

I've worked on splitting before, mostly states down to counties. When they're split too far, it makes it hard when you're trying to locate a specific site if you're not sure what county (in this case town) it's in. It was especially difficult when working to fix links to the original list that after splitting needed to be linked to the new more specific list. I also prefer geographic groupings as a user when traveling. The fewer lists I have to figure out that I need and print before I travel, the better, and the less likely I am to miss something. I would split into three geographic groupings. I guess Markvs88 will probably think these arbitrary, too, but they make sense to me. And, all lists are arbitrary. A lists is of whatever you define it to be of. So, as long as there's consensus, how you (or the wiki we) choose to define a list is not arbitrary.


 * National Register of Historic Places listings in western Hartford County, Connecticut - I would split this along a north-south line defined by the eastern borders of the towns that include parts of the Farmington Canal-New Haven and Northampton Canal, from Suffield south to Southington, and including the towns west of them. This section would include 119 places, one of the three largest towns which would otherwise be split by itself and have the added benefit of making it unnecessary to list the canal in five separate lists.
 * National Register of Historic Places listings in southeastern Hartford County, Connecticut - Continuing the division already begun when Hartford was split out, split the eastern section along the northern border of West Hartford, Hartford, East Hartford and Manchester. The area south of these borders, exclusive of Hartford would have 104 places, including one of the three largest towns which would otherwise be split by itself.
 * National Register of Historic Places listings in northeastern Hartford County, Connecticut - The area north of Hartford, etc., would have 66 places, including one of the three largest towns which would otherwise be split by itself.

I'm not hugely invested in this, so I'll not fight it if you decide to go with the town specific splits. I just thought I'd share my viewpoint as an editor and user. For town splits, I would prefer the >15 compromise outlined by SPhilbrick above. It would actually include one more split out article than he mentioned, Simsbury with 16, leaving 142 for the remainder list. This would split this one list into 6, instead of 11 with the >10 option and reduce the mentions of the canal to three lists, from five. Lvklock (talk) 02:44, 22 September 2010 (UTC)
 * I missed Simsbury, ironic as I took some of those pictures. I've updated the list. -- SPhilbrick  T  11:41, 22 September 2010 (UTC)


 * Oppose. To me, this is basically the third option. Let me put it this way: what would "National Register of Historic Places listings in northeastern Windham County, Connecticut" look like? Or National Register of Historic Places listings in southeastern Middlesex County, Connecticut? If we're going to come up with some sort of definitions based on something nonexistant (counties), they shouldn't be variable too, IMO. Markvs88 (talk) 03:01, 22 September 2010 (UTC)

Not surprised that you didn't like it. I didn't think you would. But, I think your argument is specious. My suggestion is a specific solution for the specific problem of how to split this specific county. Though I truly don't care which way you as a group choose, I do kind of resent your off the cuff refusal to even consider it. You've not presented any solution to the problems that have been mentioned with oversplitting. Basically you just don't like it, so it must be wrong. I have no idea what the split for Windham County would be. Does Windham County even need to be split? No, it doesn't, so that's a pointless question. Middlesex County doesn't need to be split either, but if it did you'd split Mddletown out before even trying to come up with some other designation. The point is that in the county with CT's capital (where I'd expect a lot of history), you have the unusual condition of having to further split the county. So, you figure out some way to do it, and you do it there. It doesn't have to apply, in theory, to every county. It's silly, IMO, to refuse a solution in one place because it doesn't apply to another place that may or may not even have the problem! But, you're obviously not interested in addressing any of the issues that splitting in the way you're headed may cause for other editors and users, but instead just about making sure it's not "artificial" in some way. Your solution may be more natural, but IMO, it's less useful. I'll bow out now and leave you to it. Lvklock (talk) 03:37, 22 September 2010 (UTC)


 * How can opposing setting up "areas" that have no real world application be specious? Watch this:

Farcical proposal
I propose that we get rid of all counties and merge cities and towns according to their endings. All "fords" (Stratford, Wallingford, Hartford, et al) will be one group. Then all "fields" (Fairfield, Mansfield, Bloomfield et al). Then all the "urys" (Danbury, Glastonbury, Southbury, et al) well, you get the idea. And this idea makes about as much sense as randomly deciding which county needs to be split and into what made up components. Actually, Hartford County is no more a special case than New Haven or New London (both of whose lists are longer than Hartford's btw), as they are the other two original colonies. I would also say that it certainly should apply to every county, as these lists do grow and we'll just be back to finding another "band aid" later... or indeed for those counties now. Markvs88 (talk) 12:06, 22 September 2010 (UTC)


 * Hey, Markvs88, i don't think you have to mock another editor's suggestion, and your opposition is overly strong, IMO. There's no need to jump in with instant, bolded "Oppose" vote, frankly.  Obviously different real suggestions are being given.


 * I do kinda like the "Fords" vs. "Urys" vs. other ones! :) But, i remain more enamored of dividing by significant geographical lines, like the Connecticut River, and/or by adequately descriptive names like "southeast Hartford County" where the arbitrary but objective division is explained upfront in the article.


 * The points raised by Lvklock, with which I agree, are in favor of the list-articles being USEFUL. Useful for editors/photographers/visitors, and useful for readers.  It is one requirement of Featured Lists (see wp:FLC) that lists be useful.  Divisions by geographic grouping are clearly more useful in some respects;  divisions into spotty small towns are less useful in those respects. --doncram (talk) 01:37, 23 September 2010 (UTC)


 * Doncram - Who's mocking? That was a logical counter example (and labeled as such!). And now speaking my mind is needless? Please accord me the same latitude that you accord yourself. If the point of all this discussion is to decide something, we need to know where people stand on each alternative! Yes, lists need to be useful. Arbitrarily putting towns together is not useful by definition. There *are* no geographic groupings in Connecticut, therefore there is no reason for "us" to adopt any which would be non-verifiable. Markvs88 (talk) 11:55, 23 September 2010 (UTC)

Fifth alternative
Perhaps we could do both? I think the size and slow editing is a function of the images, not the text. What if we had a text only sortable list for the county, then break out according to one of these town based alternatives for a list including the images? I haven't fully thought this through, it has implications for updating of new properties, but perhaps worth discussing. -- SPhilbrick  T  11:49, 22 September 2010 (UTC)
 * I believe it is the parsing of the coord template that slows the loading of the page, not the images. --Polaron | Talk 12:48, 22 September 2010 (UTC)
 * That seems plausible. Maybe someone with more techy skills can run this load tool and see if the results shed any light on it. I've run it, but do not know how to save sharable results. I see the images in the table, but I didn't see the parsing of the coord template, but this is out of my area of expertise, so maybe someone else can look at it and make more sense of it.-- SPhilbrick  T  13:52, 22 September 2010 (UTC)

<-For those not interested in actually running the tool, it generates a number of warnings, one of which is: Warning! The total size of this page is 652295 bytes, which will load in 168.00 seconds on a 56Kbps modem. Consider reducing total page size to less than 100K to achieve sub 20 second response times on 56K connections. Pages over 100K exceed most attention thresholds at 56Kbps, even with feedback. Consider optimizing your site with Website Optimization Secrets, Speed Up Your Site or contacting us about our optimization services. -- SPhilbrick  T  14:00, 22 September 2010 (UTC)