SketchUp has a really annoying bug introduced in 2021 that affects all non-current MacOS versions: The Name column in the Tags window won’t resize. Finally found a solution that works on a message board:

These steps have been working for anyone who tries them:

1. Open SketchUp, and from the welcome screen choose a previous file. Name column won’t resize.
2. Close SketchUp and open it again. Start a new document. Name column can resize.
3. Open the step #1 file from the File/Open Recent menu. Name column can resize. <This doesn’t work for me. The only thing that works is to close the welcome screen, create a new file, then open from the Finder.>

SketchUp has a weird bug with some of its Inspectors. I’ve seen it most often with the Tags Inspector and if I recall correctly, with the Components Inspector, although that one hasn’t happened in awhile. In any case, what happens is that, for reasons unknown, the Name column gets narrower so you can’t read the full name of the tag or the component. I’m running Mac OS 10.14.6 (don’t ask; it’s tldr) and SketchUp 21.1.331. The only thing I’ve found that works for me is this:

1. Pull the offending Inspector out of the Inspector cluster to a place by itself on the screen.
2. Restart my Mac and.
3. Open SketchUp to an empty file. This is important. For some reason, opening a file with things drawn in it won’t work.
4. Grab the lower right handle of the Inspector and swirl it around the screen. Once you pull it as far as you can to the upper left then back out again, it should expand the Name column.
5. Put it back in his place in the Inspector cluster.

For over a decade, I tried to process and then tag every image in a shoot rated 3 stars or above with keywords. The idea was that once completed, I’d post the entire tagged 3-and-above shoot to Zenfolio, but with hundreds of thousands of images, it has long been obvious that the task will still be mostly incomplete after I’m gone.

Rebuilding the Original Green site has made me go back and process a few thousand images so far, and also rethink my workflow. When I started rebuilding the site, I began two folders of web-optimized images: “presos 3840x2160” and “tweets 1920x1080.” In other words, 4K images for presentations and HD 1080 images for social media. And I only use images I’ve rated 4- or 5-star, so it’s my best work.

There was a problem. I only learned recently that Photoshop’s Save for Web will allow me to save the keywords and other metadata to the web-optimized files. So now I can save all that to all future images, and could technically go back and do so with the 1,100 images already saved with no metadata, but there’s another chicken-and-egg problem:

Because I’ll probably be gone before my hundreds of thousands of pics get tagged with keywords, very few of them now have any keywords beyond the basic location set which is continent>nation>state/province>city>place within the city, which I do with every shoot. So if I process a random 4-star pic, that’s all I’m likely to get.

Here’s what I just started doing this morning: going through a shoot in Photo Mechanic, if I see a 4-star or 5-star I need for something, I first go into the metadata and write a caption, which should include some of the keywords I’ll eventually tag photos in the whole shoot with someday. So even though it’s not the whole keyword set, it’s good stuff to search for.

Once captioned, I do the normal things of opening with Photoshop’s RAW converter, saving a jpg in the shoot folder, then saving the press and tweet versions. And if I need the pic for originalgreen.org, I save that version as well, which is either 2560x1440 or 720 square. Now, everything carries the caption with it.

The caption has other uses as well. I usually tweet something about these images since they’re my best work, so the caption becomes the tweet, the caption on Instagram, and the beginning of a Facebook post on the image, depending on where I want to put it. I also use it for alt text for images I’m using on my websites.

Whaddaya think?

This image was my first work to go viral, thanks to a repost by Lloyd Alter. I’ve long intended to do more of these, and finally started back this morning. Here’s my current workflow:

1. I work from Google Earth Pro; be sure View>Scale Legend is checked. Navigate to the site you want to use, then hold down the R key to rotate the view to straight down instead of slightly 3D.

2. Do a screenshot by marquee, making sure to include the scale legend in the lower left.

3. Create a new file in Photoshop from the screenshot on the clipboard. Save as “<place name> 1 original screenshot.jpg” within the folder “<place name>” and close the file.

4. Duplicate the file in the Finder & open.

5. Select All, Copy, and Paste. This creates Layer 1 where all the work is done (more on this later).

6. This step may require a bit of manipulation. With the rectangular marquee, draw a rectangle with the left side of the rectangle precisely on the left end of the scale legend and the right side of the marquee precisely on the right side of the scale legend. Write down the numbers. In this case, 325′ (the number on the scale legend) = 984 pixels.

7. 325 x 3 is 975, which is really close to 984, and it would be really convenient to scale the image so that 1 foot = 3 pixels, so I scaled it down by a proportion of 975/984. Something was slightly off, likely due to the fractional proportion, so I tweaked it slightly until the 325 foot scale legend was precisely 975 pixels wide. I’ve decided to always use a scale where 1 foot = a whole number of pixels to allow for accurate scaling between images to be compared.

8. Save the scaled file as “<place name> 2 1′ = 3 pixels.jpg” or whatever the scale actually is.

9. Duplicate the scaled file for trimming. My intent is to trim to the “essential edge” of the place. By that, I mean a clean edge of some sort like the inside of a surrounding ring road, the inside of a body of water, the outside of a city wall, etc.

10. Begin by using the Polygonal Lasso tool in Photoshop to do most of the trimming, cutting to all clean edges and just outside the edges of overlapping items like trees. Zoom in to 400% to clearly see the edge and trim only that area inside the visible window. Taking it in smaller bites not only means you can see more clearly, but also has the benefit of making mistakes easier to correct because if you go halfway around the place and then mess something up, there is more to fix.

11. Go back around and use the Eraser tool set to 7 pixels, Brush mode to make the trees look more natural, since the Polygonal Lasso tool leaves them with straight edges without thousands of tiny clicks.

12. Delete the Background layer. This leaves the place image surrounded by transparency, so I can copy and drop into another image and there’s no white file-size rectangle obscuring everything below.

13. The image originally included the scale legend in the lower left corner, so it’s likely bigger than it needs to be, so crop the image fairly tight around the remaining place image within the transparent surrounding area, but don’t get it so tight that an edge of something inadvertently gets trimmed to a tell-tale straight line.

14. Save As “<place name> 3 1′ = 3 pixels trimmed.psd”. It must be a psd rather than a jpg file at this point to preserve the transparency.

15. Be sure to keep all three files (original screenshot, scaled, and trimmed) in the same folder. It’s both a trail of breadcrumbs in case Tumblr ever lost this post, and it’s also a way to prove the scale in case someone questions it. Hundreds of people questioned the scale of the images at the top of this post when it first went viral, and while I could describe the workflow, I couldn’t actually prove what I had done because I hadn’t saved the intermediate steps.

I’m finally biting the bullet and rebuilding all my websites in @rapidweaver. I used Apple’s iWeb until they stopped supporting it about a decade ago. After much agonizing, I went with Sandvox, which was similar but better on several counts. But now they’re out of business.

Learning any new app takes time, this thread will be “notes to self” for the next site I build because things that take a week to figure out can often be explained in a minute or two, so hopefully others will find this thread useful.

A few things about my web presence: 1. I do as little coding as possible because I don’t build sites for a living. My solution must have a WYSIWYG interface. 2. I have several sites, so any solution that charges by the site and month won’t work for me.

The newest version of RapidWeaver does a great job with responsive web design on a range of devices, automatically serving up images scaled to those devices without me having to create all those image sizes for each one, like other web-building apps require. A huge timesaver.

I strongly considered Wordpress but the learning curve was just too steep to get the look & feel I wanted in a reasonable amount of time. But the Stacks on which Rapidweaver sites are built remind me a lot of the new Wordpress Gutenberg systems of blocks.

Wordpress is the world’s #1 site builder & has a huge following so it should last a long time. But the Wordpress ecosystem is a notorious hacker target since it’s so big. My RapidWeaver sites are built on my Mac and hosted by my webhost. And the RapidWeaver community is big.

To get started with @rapidweaver: 1. Buy the app at realmacsoftware.com. Stacks are the essential building blocks, which you can get from @YourHeadSupport. Next you’ll need a robust framework of components. The two top candidates are Foundry & Foundation by @weaversspace.

I started with Foundry but then went to Foundation once I discovered they play well with Hoefler web fonts, which I’ve used for years. weavers.space has a bunch of other great stacks built on Foundation, and a robust user community.

I initially bought Font Pro with the intention of using it with Foundry, but couldn’t get my web fonts working so decided to get Foundation to replace Foundry hoping they would play nice together, both being part of the weavers.space ecosystem.

I struggled for a couple more days getting my web fonts to work. Finally, I asked myself whether it might be because I hadn’t published the site yet. To that point, I’d just been building the basics on my Mac. So I published. Nothing happened at first. But then I noticed this at the bottom of the Hoefler Font Family Setup box: “In order to preview your fonts, you must add "127.0.0.1” to your domain list in Hoefler (RW7 only)“. I’m on RW8.7 but thought I’d give it a try. It worked! Not sure which fixed it, as I did one right after the other and it might have taken a minute for the publishing to trigger something in my Hoefler account, but I don’t really care since it’s working now.

weavers.space developer @joeworkman has lots of tutorial videos which are really helpful but it took me awhile to realize he’s developing stuff so fast that many of the videos are from previous versions. For example, he talks a lot about swatches but I couldn’t find anything that said "swatch” in Site Styles, which is the nerve center of Foundation. That’s because since the video was posted, swatches have matured into Font & Text, General, and Component swatches. Once I realized that, things got much easier.

I have Font & Text swatches for my text style, link style, and heading styles from h1 to h6. I have General swatches for background colors and for margins. You can use swatches as a “class,” which you can apply to almost any stack on a page by tagging it with the class name.

You can also apply swatches to a long list of stack types. Doing this means you don’t need to tag them with the class, but it affects all such objects on the page. As a result, I do mostly class swatches. I have two Component swatches: Top Bar Styles & Menu Styles.

Speaking of Top Bar & Menus, I’m rebuilding all my sites with a black background (more on that later). Some devices are small enough that the menu wraps to more than one line, and the leftover space was a light grey, which looks bad.

I worked for a couple days trying dozens of options before figuring out a hack. At the top of Site Styles, there are Component Colors and Text Colors. The latter is easy, but I don’t know what the Component Colors do because they haven’t shown up on anything I’m doing yet.

Exasperated, I did a screen shot of the grey leftover wrapped menu area and drug it around the screen, trying to see if it matched anything. One color is called light grey & looked like a perfect match. I changed it to black, and that solved the problem.

Knowing all that stuff now, I can probably set up my next site in a few minutes instead of a couple weeks. Hope this helps someone else. In any case, the rest of this thread will be about my new site design system.

I’m doing black background design for several reasons. In 1984, Macs brought computers out of the DOS black-screen world into the light of black letters on a white background, like real life. But now black-screen design is cool, including on apple.com.

Black-screen design also uses less energy, which might not be much on one computer, but is a big deal on many. And since http://originalgreen.org is about green building and living, it makes good sense. Plus, I use a lot of black in other places as well, so it’s consistent.

I started studying the best new sites in late November, and noticed several things: With responsive design being essential today so you don’t get dinged by the search engines, designs with many small pics are out, as is clutter, including the page-top banner.

The top of my page has a simple SVG (scalable vector graphic) logo on the left and a menu bar on the right. Below that is a big, beautiful image. I save a 4K (3840 x 2160) version of my best work for presentations & a 1920 x 1080 for tweets and other social media posts. The great thing about RapidWeaver is that I can use the 4K version of my images and RW will automatically serve up the right sizes to each device, including Retina devices. And by using 4K versions, they’ll still be good when sites get to 3840 pixels wide (mine are 1280 now).

Below the title image is my page title and text. @joeworkman has moved Foundation to markdown text, which I’ve never worked with before. I really like it because you can switch between text & header styles with just basic keystrokes all in the same text box using no HTML.

On long blog posts, I’ll definitely intersperse images with text throughout the post as I’ve always done, but those images will all be full-width cinema-proportion 16:9 images so they’re big and beautiful.

Something else I noticed about the best new design is that sidebars are mostly out today. The best sites take the clutter that would have been in a sidebar and put it in a box at the bottom of the page, so the user experience is just beautiful images & text.

My footer box includes a Google site search bar, links to where you can buy my book, email links on speaking & ideas to write about, a subscribe box, social share buttons & follow buttons, a Facebook Comments box on most pages & a sitemap of the 1st & 2nd tier tabs.

originalgreen.org has hundreds of pages, and the Facebook Comments box requires the URL of the page it’s on. I’m building the new site at dev.originalgreen.org but when the site goes live, I’ll have to fix the comments box on every page.

Also, all the inter-site links I’m building right now will have to be re-done once the site goes live. So I have a radical thought: I don’t currently sell anything direct from any of my sites, so I’m wondering about just building the new site on originalgreen.org.

Another great thing about RapidWeaver is that I can make URLs pretty much whatever I want, including the identical URL of the old page. So there’s no need for redirects, which are a huge task on a big site.

Building a site live at the permanent URL means that if someone comes to a new part of the site, they’ll only see the part that’s currently built, but if they have a link to an old page, then they can still see it right up until the day the page is rebuilt.

Does anyone know why this strategy wouldn’t work? I feel like I’m standing at a point of no return on this. What am I missing? Develop on dev.originalgreen.org or right on originalgreen.org? All thoughts appreciated; if I’m gonna do this, I need to start now.

I’ve been building a strong image keyword taxonomy for over a decade, but have been using it loosely. The idea is to tag images on Mouzon Images and also the Catalog of the Most-Loved Places with keywords so people can search easily for useful stuff.

But I’ve been using it very loosely until now, tagging just the stuff that I think of. And my taxonomy is much better than that, so now I’m upgrading my process to be worthy of the taxonomy. Yes, it takes quite a bit of time for a big shoot, but the images are much more useful if you can find what you’re looking for. So here are the most important things I’m doing:

Don’t Over-Tag

If I tag every detail I can see fully zoomed into the image then a search turns up so many images it’s overwhelming. And the details may be so small that they’re not high-res enough to be useful in a publication or even online. So what I do now when I begin tagging is to set the thumbnails two notches bigger than normal in Photo Mechanic (much faster than Lightroom). This results in 12 thumbnails wide on my screen, which is a 4K. If I can’t see it in a thumbnail that size (256 pixels wide), it doesn’t get tagged.

Narrow the Candidates

I’m working on my Quebec City shoot now, which includes 320 images rated 3-star or higher, so that’s a lot of images to go through. I have some keywords in the taxonomy that I haven’t used before, but it turns out they’re really useful. I tag “streetscape” which means an image with at least 3 buildings and a street because individual details are generally too small to be useful. Same for the “from high land” and “from rooftop” tags. I can easily search for these and tag the checkboxes all at once, then set Photo Mechanic to only show untagged images. As I work through the tagging, I try to find atypical images and tag them first so I can tag the checkbox and eliminate them from the image browser as well.

Two caveats: whenever there’s very little to tag on an image and I won’t need them for any of the Frameworks (Light Imprint, Original Green, Pattern Language, Sprawl Recovery, or Transect) I go ahead and do the Close Out (see below) so I don’t see them again. For everything else, narrowing the candidates is something I do for each major subcategory, like building>exterior. When I get to the next major subcategory like building>type I’ll first show all images, untag them all (Command-minus), find the closed out images, tag them, then just show untagged to begin. In other words, a clean slate of all the images I still need to see.

March Straight Through

My keyword metacategories are architecture, art, building, creature, environment, event, frameworks, issue, location, object, plant, scene, texture, thoroughfare, and transportation. After taking a pass through all the 3-stars to see which should be elevated to 4-star, then going through the 4-stars to see which should be elevated to 5-star, I begin with the scene metacategory and tag the ratings, then the aspect ratio (landscape or portrait) then the other appropriate scene tags. Then I basically march straight through the rest of the metacategories in alphabetical order, which is something I’ve never done before until now. Yes, it takes time but produces thorough results.

Search Smart

On some of the sub-subcategories like building>exterior>4 eaves & roofs>dormer, it’s faster to scroll through the thumbnails and pick out the images with dormers. On other sub-subcategories like building types, there are just too many, and the best thing to do is to open the preview window showing a single image large and scroll through, tagging each type that’s obvious. Some won’t be an obvious type, so it’s easy to go through a run of those, then select them all and check the tagged box on all of them by doing a Command-plus.

Focus On the Best

When working through a place like old Quebec City which has so many good examples of some of the frameworks like the Original Green, it is tedious to try to tag every image. So I turn off the 3-stars so that I’m only looking at the 4- and 5-stars for tagging. Why not promote the principles of the frameworks using the best images?

Close Out

In order to reduce the number of images I’m looking at, I’ve added a keyword to the scene metacategory: OK. Adding this to the keywords for an image means I’m definitely done adding keywords to that image, and can therefore easily tag all the OK images and then not see them on any new pass.

Video

I now shoot with my iPhone 11 Pro, which simplifies things greatly. When shooting a presentation where I’m stationery, I use a mini pod I carry in my backpack. If possible, I sit right behind the projector and set up the mini pod on the projector, assuring me that there will be a clear view of the screen & the presenter. It also is a good bet that I’ll be able to find an outlet so I can keep the phone charged.

Audio

With a camera as good as this iPhone, the weak link is usually audio, although my iPhone X did better audio than my $150 Rode shotgun mic on my Nikon D610, which some consider a professional setup. And the iPhone 11 Pro is better, but the big problem for me is shooting outdoors and getting all the terrible wind noise on a breezy day. So I’ve just ordered a Shure MV88 iPhone mic, which some regard as the best, and a Movo “dead cat” windscreen. I’ll post later how this works out.

Gimbal

I have an Osmo Mobile 2 gimbal, which is highly rated for good reason. It’s just persnickety because you have to do everything in just the right sequence. I’m doing this workflow in large part to document that sequence so I can look it up later:

1. Be sure to keep it charged, as it takes easily 2 hours to fully charge.
2. Attach the phone with the gimbal turned off because you’ll need to balance it and it’ll fight you if it’s on, and you won’t get a true balance. Follow the instructions, which basically tell you to adjust the phone with it hanging with the face looking straight down at the floor so it’s flat top-to-bottom and side-to-side.
3. Unlock your phone.
4. Power the gimbal on by holding the power button for several seconds until it snaps to attention.
5. Here’s a tricky part: we actually have two Osmos, one for me and one for Wanda, as she has begun filming with me at CNU each year. I thought I could pair either one with my phone, so I began trying a few days ago to get one of them to work, but the phone would never recognize it. Finally, I decided to try the other one and got it to work. Apparently the gimbal has retained a memory of my phone or vice versa, even though this is the first time I’ve used the gimbal with this phone. In any case, go into Bluetooth on the phone and make sure the gimbal’s connected. Then go to the DJI GO app which is the software of the Osmo and so long Bluetooth is seeing your gimbal, you should be able to click on Connect Your Device after the opening screen goes away.
6. Once it’s connected, it opens the DJI camera window on your phone. This is important: to get full functionality, shoot from here, not from the default Camera app. The red button on the gimbal is the shutter button; on the left of the screen you’ll see a toggle between video & single shots. The zoom button is on the left, so a right-handed person can easily do everything with just their right hand.
7. Once you’re done shooting, you need to get the video & single shots off DJI and into Photos. To do this, first power off the gimbal and take your phone out. Then go back to the DJI app and go to Editor, at the bottom menu bar. Click on each video, then the download icon in the lower left corner. Click Save video to your device, which puts the video in Photos. Now you’re ready to move on to post-production.

Post-Production

Here’s how I do post-production on the video itself. I do everything on a Mac, so all the terms used here are Mac terms, and I’ve capitalized them for clarity.

1. I have an Album in Photos called Movies which filters out everything except movies, so it makes it really easy to focus. This is very important because if I had to sort through the many thousands of photos to find the movies, I would surely miss some of them.
2. Select the new clips to bring into iMovie and drag them to the Desktop. This will make a copy on the Desktop but leave the original in Photos. if you try to drag them directly into iMovie it’ll just get the opening frame as a still photo. I don’t know why; that’s just the way it works. IMPORTANT: for some reason, if you try to drag 30 or more clips, it can take close to a half-hour to complete, so only drag 3-5 at once and you’ll get done much faster. For some reason, dragging a lot at once chokes it down.
3. Open iMovie then Open Library to select the Library you want to work in. Because movies take up so much space, I do all my movie work on an external 4TB LaCie Rugged drive. I have two libraries on this drive: one for movies I’m working on, and the other for the completed movies.
4. In iMovie there are two main windows: Media & Projects. Toggle between them at the very top. To bring in new clips, click the Media button.
5. The far left column of the iMovie window is the Libraries window. Click on the current year. Then File>New Event or Option-N. Name this event for the clips you’re about to bring in.
6. Drag the movies on the Desktop into the new Event.
7. Click Projects, then click the template you want to use and duplicate it and rename it for the new movie. I keep a template for every type of movie I create, so I don’t have to do it from scratch every time.
8. Open the new Project by double-clicking, which takes you to the project window where you’ll build the new movie. Drag the clips you want into the Project, being sure to “snug them up” to the template’s opening clip so the cross dissolve between the template’s opening & closing clip goes between the clips you’re dragging in and the closing clip.
9. I have a 4-second opening title and a 5.5-second closing title, which begins in the last 1.5 seconds of the last imported clip then fades to a Gradient screen in the last 4 seconds. The opening title is in the template over the template’s opening clip, but that’s just a placeholder because the title is meant to start at the beginning of the first clip, so drag the title to the beginning of the first imported clip and delete the template’s opening clip.
10. If there’s more than one imported clip, insert transitions between them. I use Cross Dissolve because it’s simple & doesn’t call attention to itself because my videos are all about content.
11. Trim clips for a smooth transition. To do this, you’ll need to find the little slider at the center right of the window just left of Settings and slide it pretty much all the way out to the right to stretch the clips so you can trim to exactly the right place. This used be a bit tedious when filming a presentation with my Nikon D610 because it would only let you get 20 minutes of video at a time, so I’d have to edit back to the last complete sentence the speaker said on the early side of the transition, then to the beginning of the first complete sentence on the late side. Now, however, because the iPhone has no such limitation (except for total storage space) I don’t have to worry about that anymore on a single event. So the only transitions I have are when I film several different clips and stitch them together. In that case, if I’m explaining something, I just be sure to leave a second or two of silence at the beginning and the end, which makes it really easy to trim properly. To do precise 
12. Click on each clip, then make modifications. I used to use the Cartoon clip filter because it makes the images soft, vibrant, and romantic, but a lot of people complained so I’ve decided to use no clip filters at all. Now, I only make two modifications: under the Volume icon I click Auto, and under the Equalizer icon I click Reduce Background Noise and leave it set to the default (50%) and select Voice Enhance under Equalizer.
13. The movie should now be ready to share. But sharing is a space-hogging process, so before you do, make sure you’ve deleted all the videos you copied to the Desktop, and empty the trash can. If the share fails, you might even restart your Mac and then only open iMovie. I use Clean My Mac, and run that before sharing as well.
14. There’s an option in File>Share to share directly to YouTube, but don’t do that. It always fails. It seems that really short clips (like my 4-second closing clip) might be the problem. So I share to a file and then upload the file to YouTube (more on that later) which has the added benefit of letting me share that file directly to any other site that might not play well with YouTube, either now or in the future.
15. In File>Share I write a Description that includes not only the story I’m telling, but also where the video was shot in case someone is searching for videos on a place or building. I give it one Tag such as “urbanism” because tagging is easier in YouTube. Format is Video and Audio, Resolution is 1080p (which is HD), Quality is High, and Compress is Faster.
16. Click Next, which will take you to a window where you tell iMovie where to put the movie. On my external drive, I have two folders: movies to upload, and uploaded movies. In the movies to upload folder, create a new folder with the name of the movie and click Save.
17. My title fades in and out at the beginning which is a nice effect, but that means the first frame has no title. Thankfully, YouTube allows you to assign a “thumbnail.” So once the movie has been created, open it (it’ll come up in QuickTime Player) and run off a second or so until the title has fully faded in. Edit>Copy, go to Photoshop, create a new file the size of the clipboard, and paste. Merge Down (Command-E) then Save for Web (Command-Shift-Option-S) in the same folder as the movie using the name of the movie plus title, so the Civic Gift to the Street movie would have a title shot named Civic-Gift-to-the-Street-title.jpg.

YouTube

You’re now ready to post to YouTube, but first a bit of background. I don’t know why Google makes things so difficult to find, so I’m adding this to remind myself how to get back to it whenever I have long gaps between posting videos. Three years ago, I consolidated all my Google stuff under one email address when they finally created “Brand Accounts.” Previously, I needed one email address for my Original Green content, another for Mouzon Design content, etc. Now, I’m putting up new videos on the Original Green channel but spent hours day before yesterday unsuccessfully trying to figure out how to actually manage the channel rather than just viewing it. I just discovered what I was missing: Under my avatar (my head shot) for my Google account, if I pull down to Switch Accounts, it’ll let me choose any of my other brand accounts. I had seen it but assumed it meant go to an entirely different accounts, like when I had multiple accounts on different email addresses. Once I’m signed in as the Original Green Brand on the Original Green Channel, I do the following:

1. Go to studio.youtube.com.
2. Click Upload Video and select your movie file. This will open the window where you can add details about the file. For some reason, YouTube doesn’t pick up the description I wrote in the Share step, so I have to rewrite it.
3. Just below that is a frame you can click to upload your thumbnail, or title image as it’s called above.
4. Next is the Playlists window, where you can add it to an existing Playlist or create a new one.
5. Under Audience, click No, it’s not made for kids. Then under Advanced, do the following:
6. No, don’t restrict my video to viewers over 18 only.
7. I don’t do paid promotion.
8. Tags can help people find videos, so add the obvious ones, including the title of the video plus others like “urbanism,” “New Urbanism,” etc.
9. Enter the recording date & location.
10. The category of all my videos is Howto & Style.
11. I don’t currently add end screens or cards because I don’t know anything about them.
12. On the next screen I Publish as Public.

Repost & Promote

That’s it! Except to post the video to Twitter, Facebook, and your other New Media nodes.

This is not quite complete; I’ll update the last item or two later.

I finally got my entire tweet catalog organized, hash tagged for topics, tweet casted events & speakers earlier this year. I also established a Fave field for what I consider good tweets and a Classic field for the great ones. Other fields include date & time (obviously), a checked/unchecked field I use for general organizational stuff, a field that starts with the text of the tweet followed by the speaker’s name if it’s not me, and a field that counts the number of characters in the tweet. This will be the first time I’ve added to the organized catalog, so I’m recording the steps so I won’t have to figure it out again next year.

Request Twitter Archive

The first step is to get my Twitter archive. In a web browser on my laptop, there’s a More button at the bottom of the left sidebar. Click that, then on the pop-up window click Settings & Privacy. On the next screen, click Your Twitter Data under Data & Permissions. Under Download your Twitter data, there’s one button for Twitter and another for Periscope. Click Twitter. It’ll ask for my password. It takes awhile to put it all together, but in the next hour or two I’ll get a notification saying it’s ready. Now, Twitter lets people download their data only once per month, so it’s important to pick the time carefully to avoid copying a bunch of new tweets into the catalog by hand.

Convert Archive

Until this year, the archive included a CSV of all tweets back to the beginning. This year, they added a bunch of new stuff, including all of the media (pics & movies) you’ve ever posted, which is nice, but they inexplicably left out the Index.html file where you could see all the tweets just as they appeared and the CSV file which you could easily open in any spreadsheet. Now, everything except the media is in JSON files, which only a tiny fraction of people know how to manipulate. It took me several days to figure out how to convert it so I can open it in Mac Numbers.

Searching for a solution, it’s obvious many other people have had the same problem, and none of the ones I found had discovered a solution that non-programmers could easily do. So I tried to find an app, or find an online converter, which you would think would be easy. But every one I tried had some sort of problem. They either said “conversion failed,” or they tried to extort me for hundreds of dollars per year to get what they said was a successfully converted file. Obviously, I’m not going to do that.

Finally, I noticed that several of the failed conversions kept saying there was a problem with the first line of the file. Fortunately, I needed to download my archive a few months ago, and was able to convert it successfully on one of the online converters. It turns out that Firefox can open Javascript files as text, so I opened both of them. Sure enough, the first line of the new file that didn’t work was “window.YTD.tweet.part0 = [ {“ while the first line of the old file that did work was just “[ {“. So I deleted the offending “window.YTD.tweet.part0 = “. And I should note that while YTD implies that it’s only Year To Date, the file actually contains all my tweets back to the beginning like it always has.

The converter I originally used, and which I was really happy with, is the very simple & totally free https://onlinecsvtools.com/convert-json-to-csv by browserling. Click Import From File on the JSON side and it’ll load the file in a few seconds (upload time depends on file size, of course). Click Use Headers in the box just below the JSON box so it’ll convert the first line to a header row. Very quickly, it converts the file and the data pops up in the SCV window. If it doesn’t pop up, there’s something wrong with the file. Click Save As… to get the file. Unfortunately, it only read a few of the columns. While they were the most important ones, there was a lot of useful stuff missing.

So I went back and looked at a conversion I’d done in early November, and it was on a site convertcsv.com. So I went back to that site to convert the data. Simply put, it was one of the most painful things I’ve ever done. Nearly all day New Years Eve was spent trying to shepherd it through. For the first several times, the browser kicked cranked on it for several minutes, then reset itself due to lack of memory. And to be fair, xfinity was incredibly slow that day, so that may have had something to do with it. I ended up doing it in Firefox, since I had a bunch of windows open in Safari and didn’t want to close them. Before running it, I restarted my Mac just to be sure it was as clean as possible. Still, it repeatedly ran low on memory, so I’d follow it in the CleanMyMac widget in the toolbar, and free up memory every time it got below 10 gigs. It was absolutely painstaking because I couldn’t do anything else other than sit there and play games on my phone and watch it out of the corner of my eye. But fiiiiiiiinally, it finished and I was able to download the converted data.

Delete Columns

When I opened the file in Numbers, I got an error message saying that Numbers only reads the first 256 columns, and there were over 500. Assuming that Twitter probably put the most important stuff near the beginning, I said OK and opened it. Turns out that was correct. In the spreadsheet structure, they put the first few of an item type near the beginning, then I’d find more buried deeper. Towards the end, data was getting really sparse so I doubt there was anything really essential in the columns that didn’t load. And I found many columns that contain stuff I know I don’t need, or that duplicated other info, such as the full URL of a link (which I kept) versus various other iterations (like the Twitter-shortened URL) of exactly the same thing.

Without getting into the rationale of why I deleted columns, here are the ones I kept: favorite_count, retweet_count, created_at, full_text, entities/urls/0/expanded_url, entities/media/0/expanded_url, people mentioned & their usernames 11 deep (I’ll probably only need the first one, if even that), and hashtags 6 deep.

Fix Date

1. There are several extraneous things in the date & time, so I create a new column where I’ll create a text formula to keep only what I need. Currently, the date & time look like this: Mon Jan 26 17:31:04 +0000 2009. I don’t need the day of the week or the +0000, but I do need the year with the month & date so Numbers will read it correctly. The formula is: MID(E2,5,3)&“ ”&MID(E2,9,2)&“ ”&RIGHT(E2,4)&“ ”&MID(E2,12,8). This converts the date & time above into: Jan 26 2009 17:31:04. I paste this into every cell of the column.
2. So the dates look right on the screen, but they’re still just formulas, so I create another column named date & time beside the formula column. I copy the formula column and Paste Formula Results into the date & time column, so Numbers sees it as actual dates & times, not formulas.
3. Change the date & time format to look like this: 2009-01-26 17:31:04, and so that it left-justifies.
4. Sort by date & time.
5. Delete the created_at column & the date working column.

Copy Orphan Tweets

I do a holiday mailing each year that contains some of my best work on Twitter for the year, so I have to harvest the archive before the year is over. This means there are always some orphan tweets that didn’t make it into the catalog already. So I find those and copy the date & time and the text of the tweet.

Delete Previous Years

I make a Filter for all dates before 1/1/19 (or whatever the current year is, then delete all rows remaining, as they are in previous years.

Total Faves & RTs & Measure Length

I added a column after the fave & RT columns totaling the two, for a sense of the impact of the tweet, in the interest of assigning fave & classic settings in the final document. I also added a length column before the text of the tweets measuring total characters.

Delete Short Tweets

I sort the Length column, and delete most of the short tweets which tend to be stuff like “Thanks!” For this year, the shortest tweet I found that was useful standing alone was 45 characters.

Delete Original Green Daily

I do a paper.li newsletter every day. When paper.li puts it out, it tweets The Original Green Daily is Out! with a link. So I sort by text, which puts them all together so I can delete all 365 of them at once. DON’T filter for that text, because while it’ll only show the 365, if you select those rows and delete, it’ll delete everything else in between, basically wiping out your database. If you do, Undo before you save!

Clean Out Long Copy Chains

The longest tweets will be those with long copy chains, where many people are in the conversation. My longest one this year was over 900 characters. So strip out all the usernames to see what’s actually being said at the end. There’s no easy way to do this; allow 20-30 minutes.

The Long Edit Slog

This is the most time-consuming part except maybe the hashtagging; allow a couple days depending on how much you’re starting with. For this year, after deleting the stuff noted above, I had about 3,300 tweets. After this edit, I have about 2,300 so this pass takes out almost a quarter of the tweets. There are four things to do in this pass: 1. edit the tweets, 2. assign them fave & classic status, 3. move the names of people I’m quoting to the @ column, and 4. tag the event I’m tweetcalting in the tweetcast column. The edit involves several things: 1. Because the tweets stand alone in the catalog, I strip out all the @usernames at the beginning that didn’t get taken out in the long copy chain cleaning above, and remove remnants of conversational text, leaving just the statement of principle. For tweetcasts, I move the speaker name to the @ column (either as @username if they’re on Twitter or #FirstLast if not) and the event to the tweetcast column. Obviously, once I have one of each, it’s easy to duplicate. Then I strip the speaker & event name out of the text. I also delete a lot of tweets in this step that don’t belong in the catalog; some because they’re banal, others because they’re conversational and no part of them can be properly understood standalone. Until this morning, I was deleting truncated retweets as well but may have found a way around that, so will leave them for the moment (see The Retweet Problem). Also, I don’t want to have to read a tweet more than once, so while I’m here, I assign fave & classic status. Faves can be assigned for a number of reasons: 1. compelling content, 2. a memorable event to me (it’s my catalog, after all), and 3. the number of times they were favorited or retweeted on Twitter. Classics are much rarer, and must be both unusually compelling in some way and also evergreen. In other words, they won’t go out of date quickly. Every classic is also a fave.

Column Gymnastics

Because this is a working file, I toss the columns around a lot, putting what I need side by side as much as possible. I also hide columns I don’t need to see for clarity.

Fave & Classic Cleanup

Assign Hashtags

Once the edits are complete, it’s time to assign hashtags. I’ve harvested 6 columns of hashtags from the Twitter  archive (/0/-/5/) but some of them were used to tag speakers who weren’t on Twitter or the names of events I tweetcasted. I delete these entries from the six columns for clarity, then move the columns just to the left of the text column.

The Retweet Problem

Retweets in the archive are truncated to 140 characters, which makes all but the shortest text-only retweets useless. I’ve been searching for days to figure out a way around this, and finally found something this morning. Apparently, something called Python can be used to access the Twitter API and get whatever info I want. I had an app called Komodo that I got for reasons I don’t recall several years ago, but apparently it can edit Python scripts, so I’ve downloaded the latest version. I’ve also just applied with Twitter as a developer, so if they approve me I’ll update this on the adventure.

Copy to Yearly Template

I have a template file which looks just like the big catalog file except it has 3 columns at the far left: faves, RTs & total. This is where I paste that date from the archive sheet, but which won’t get copied to the main sheet once I’ve set fave & classic designations. To be sure, the cells have a grey background. It’s important to get the data into the template at this point so everything is properly formatted from this point on.

I’ve just updated my photo keyword taxonomy and have posted it on Mouzon Design. It’s a 96-page PDF of the way I arrange my world.

I have hundreds of thousands of images on my Mac, so if I want to do a search for images during a certain time window, the search will spin endlessly after I enter the start date for the window. It was really frustrating until I realized that there’s a really simple workaround: first, search for a file name you know will turn up very few finds. Then, enter the start date and beginning date and it’ll narrow the finds down to those dates almost instantly. Then, add a new search criteria for file type and select Images. Finally, take out the file name you entered originally and it’ll find the images you need within the time window quickly.

A couple weeks ago, it appeared that Amazon had killed SellerCentral customer service from humans for anyone with an Individual account. But I found another way in that the commercial account people can’t see. Here’s how it works:

1. In SellerCentral, click Help. That takes me to the dreaded new page that seems to be a black hole. Typing something in the box and then closing it does not good for me, apparently because of my account status. But my problem is one with FBA, so under Top Solutions in the right sidebar, I click FBA.
2. That takes me to the FBA submenu. I click “Get help creating a shipment.”
3. That brings up a Manage your FBA Shipments sub-submenu with four radio button choices, but lo and behold down below that, there’s text saying “Still need help?” and “Contact Amazon” links to https://sellercentral.amazon.com/cu/contact-us? the commercial account person posted earlier. From there (and I didn’t see this from just following the link… guess because I didn’t navigate to it from my account?) I click on Other FBA Issue.
4. That brings up a pop-up from which I can actually open a case.
5. When I finished describing the problem, there’s a choice to request either an email or call, so I chose the call. It came within seconds. So all the functionality is still there, including for us lowly Individual Account people! They just didn’t tell us how to find it. Hope this helps everyone who sells on Amazon.

Ever notice how SketchUp never lets you zoom in close enough when getting a location snapshot to get a really high-res one? It just occurred to me that the true North on SketchUp should be identical to Google Earth, so I got zoomed in and got a high-res screenshot in Google Earth, opened it in Photoshop, saved it as a jpg, and imported it into SketchUp. I placed it on its own layer and resized it to match the location snapshot so I can now draw precisely to the facts on the ground instead of having to guess with the really low-res snapshot.

In fairness, my web hosting company A2 Hosting didn’t know it either, and I’ve been insanely happy with them and Apple for many years. Here’s the problem:

I’ve been on POP forever, where email is stored on my computer instead of of the server. My thought was that it was more secure here, and I could back it up. But a meltdown too complicated to explain recently made me change my mind, so I migrated over to IMAP, where anything I want is stored on the server and only archives (in my case) are saved locally.

Immediately after migrating, a strange thing happened. I use Rules in Mail to automatically move certain emails to specific mailboxes (folders) as soon as I receive them. All mail from the New Urban Guild listserv, for example, goes to the New Urban Guild mailbox on my Mac. But as soon as I changed over to IMAP, 75-80% of emails with attachments started showing up with attachments that showed the default icon of a jpg, pdf, or whatever with an empty progress bar superimposed and “Downloading…” below the filename.

I didn’t isolate it completely, but it seemed to have something to do with file size. But not precisely. So I called my web host. After an hour on the phone, it was clear they didn’t have a clue, and I had to get back to work. Their track record with me has been stellar almost all the time, but they struck out with me on this. But I left with the impression that when the email was pulled off the server, the attachment had to be pulled off pretty quickly as well, otherwise it would get dis attached from the email to wander as an orphan forever in the ethosphere.

Today, I called Apple during my lunch hour. Apple generally has the best tech support people on the planet, but even after spending 20 minutes with the senior advisor after the first guy struck out, it was clear that she didn’t know. But at least she told me how to create mailboxes on the server.

When you convert an account to IMAP, Mail creates a line for each account you have just below the On My Mac mailboxes… so it’s at the bottom of the Mailboxes sidebar to the left.So I recreated a leaner version of all those On My Mac mailboxes for only the accounts I use today, and then I changed all my Rules to redirect all traffic that would have gone automatically to the old On My Mac mailboxes to the new mailboxes on the server. And so far, everything has worked perfectly.

My big question: why did everyone from Apple to A2 fail to know this?

I got a new MacBook Pro and migrated from my quite new previous MBP, then migrated Wanda from her really ancient MBP to my previous one. Unfortunately, I was distracted after the migration when it came time to sign in to Dropbox and let her sign in as me. I was at work mid-morning when I started getting notifications that I was updating folders that shouldn’t be there.

I quickly had her pause syncing until I got home this evening. As it turns out, even with a business account, you’re only allowed one work account and one personal account, so when I signed me out, I was left with an unattached Dropbox folder. When I signed back in on Wanda’s MBP as Wanda, Dropbox renamed her current Dropbox folder as (Old) and started to build the new one.

Problem is, as much as I love our little love-nest on the 10th floor of an old office building converted to lofts, the internet really sucks. 230 gigs of data would take probably a week to transfer, and we’re planning to leave tomorrow on what could be an epic anniversary road trip in New England.

I quickly Googled, but didn’t find any solutions; Dropbox straight out says that it can’t be done. But with Dropbox paused, how would it know when I started it back up if I just took all the old stuff in the (Old) folder and moved it to the new folder?

It turns out that it didn’t know. I moved it all over and it started syncing again. There were over 135,000 files and the number of files to sync rose well above that, then up to about 270,000… but finally, the number started decreasing. I kept a close eye on the total size of the new Dropbox folder, but it has never budged. So the double-the-number files to sync apparently was the number of files on the cloud plus the number of local files, which would be right. So an apparent sync of a week or so turned into just a few hours. Wish this information was out there more broadly. What do you think? Did you know this?

I have two image banks, one of which I carry on my travels, while the other one waits at home. Each is a little over 3 terabytes, so that’s a lot of images. When I need to do work on a set of images as I travel, I can’t sync to the home bank right then, and so there are eventually a lot of images where one set has been edited to a greater degree than the other.

I eventually get them synced up, but it’s often months in between. I don’t know why, but I always go back to one of my duplicate detector apps to figure out what’s most current. But that’s not the right tool. Instead, it’s a nifty little app called Compare Folders.

It’s blazingly fast, able to compare those two 3+ TB folders in 20 seconds or so, making it really easy to see what the changes are, and what needs to be updated. Not sure why I don’t think of that in the beginning of a sync session!