irdc

Not my video. From the description:

Case: Raw Aluminum Keycaps: Unicomp Red Cyrillic on Brilliant White + Gray Modifiers Layout: HHKB-style split right shift and regular 2-unit wide backspace Configuration: Floss Modded, lubed spacebar wire, lubed keycap stems

My first (or second) experience with non-MX style switches in a keyboard. The IBM Model F is coveted as one of the greats when it comes to "mechanical" switches. This reproduction unit by Ellipse of Model F Labs serves to bring back the magic of the Model F keyboard with a lot of modern enhancements. Suffice it to say, I am very happy with the experience of typing on this board minus one small caveat. My biggest gripe with this board, and consequently any other buckling spring keyboard is the fact that there is simply no way to flip the spacebar. The asymmetrical design of the keycap stem paired with the fact that the spacebar wire attaches directly to only one side of the spacebar keycap makes it so that you are required to use this in its standard spacebar configuration. I find this to be really uncomfortable because my thumb is essentially pressing right against the hard edge of the keycap rather than the smooth edge afforded by a flipped spacebar. This was such an issue that I attempted to model my own keycap and have it resin 3D printed by JLCPCB, but to no avail. The tolerances on buckling-spring stems are really specific which makes the mechanism really difficult to replicate. I dont have the time or money to fine tune this problem. Despite that glaring issue, I still have a great time typing on this board and I find the typing experience to be very pleasant otherwise. The deep clicks are much preferable to the high pitch clicks youd often hear in MX-Style switches. The mechanism is also reasonably smooth if you take the time to lube the keycap stems with Krytox 205g0. You might also notice that there is no coating on this keyboard. When I had originally bought this board from a seller off Mechmarket, the original coating on it was a powder-coated black. I wasn't thrilled with the look and feel of it so I took it upon myself to strip the coating by submerging the pieces in a vat of Klear-Strip for roughly 2 days. Overall, I think the result turned out great and gives a raw look to the design. So far the aluminum hasn't shown any clear sings of tarnishing but I do anticipate it to eventually show some wear. There's still a marginal amount of residue left over from the stripping so the finish isn't perfect, but I think it gives the board a lot of character. Despite there being no "integrated" weight in this board like you see in a lot of premium customs, this still out-weighs my Sharkeneko 1600g to 1470g. The steel inner assembly paired with the thick aluminum chassis makes this a weapon of a keyboard.

I'm looking into migrating (part of) my ever-growing pictrs data to something cheaper than the VPS disk it's currently on. Ideally I'd like to use minio's Object Tiering to migrate object to/from cheaper storage.

Anybody using Backblaze's cloud storage product? What other options are out there and what potential pitfalls should I be aware of?

Terwijl de prik-angstige schapen doodsbang achter enkele boe-roepers aanlopen haalt de rest van de wereld inmiddels redelijk de schouders op over de pandemie. Toch wel een vonst, dat mRNA 💉💪🏻👍🏻

I’ve ordered a classic FSSK but now I'm thinking why stop at just one?

Anyone here who has a Turing Pi?

It's still a beautiful series to watch, but by now it's getting clear that season 2 is tonally completely different from season 1: far less ponderous, and funnier too.

Your thoughts? Is Brother Day ever going to get married? Where did this Mule come from? Who is Hober Mallow? And how did Poly Verisof live to such an age on a diet of alcohol and Vordaline dust?

As the title says, I'm looking for a Linux/macOS-compatible EPROM programmer. I'd like to be able to program parallel (E)EPROMs in the 27xxx range. The newer XGecu programmers seem to not be compatible with Linux/macOS, and the selection of TL866(II) clones is too overwhelming for me to be able to choose one.

Any suggestions?

I’ve recently switched from a self-built Lemmy without pict-rs to the standard docker image that has pict-rs included. However, I’m noticing that the amount of pictures pict-rs stores on my instance is quite high. Any ideas? Is Lemmy preemptively mirroring all images in all posts?

Sad that updating your Lemmy instance to 0.18.1 broke your post and comment scores? Here's a small SQL script to fix them:

 MERGE INTO "person_aggregates" AS "d"
 USING (SELECT "m"."id" AS "person_id"
             , coalesce("p"."post_count", 0) AS "post_count"
             , coalesce("p"."post_score", 0) AS "post_score"
             , coalesce("c"."comment_count", 0) AS "comment_count"
             , coalesce("c"."comment_score", 0) AS "comment_score"
          FROM "person" AS "m"
          LEFT JOIN (SELECT "p"."creator_id"
                          , count(distinct "p"."id") AS "post_count"
                          , sum("l"."score") AS "post_score"
                       FROM "post" AS "p"
                       LEFT JOIN "post_like" AS "l"
                         ON "l"."post_id" = "p"."id"
                      WHERE NOT "p"."removed"
                        AND NOT "p"."deleted"
                        AND "l"."person_id" <> "p"."creator_id"
                      GROUP BY "p"."creator_id") AS "p"
            ON "p"."creator_id" = "m"."id"
          LEFT JOIN (SELECT "c"."creator_id"
                          , count(distinct "c"."id") AS "comment_count"
                          , sum("l"."score") AS "comment_score"
                       FROM "comment" AS "c"
                       LEFT JOIN "comment_like" AS "l"
                         ON "l"."comment_id" = "c"."id"
                      WHERE NOT "c"."removed"
                        AND NOT "c"."deleted"
                        AND "l"."person_id" <> "c"."creator_id"
                      GROUP BY "c"."creator_id") AS "c"
            ON "c"."creator_id" = "m"."id"
         ORDER BY "m"."id") AS "s"
    ON "s"."person_id" = "d"."person_id"
  WHEN MATCHED AND ("d"."post_count" <> "s"."post_count" OR
                    "d"."post_score" <> "s"."post_score" OR
                    "d"."comment_count" <> "s"."comment_count" OR
                    "d"."comment_score" <> "s"."comment_score")
  THEN UPDATE
          SET "post_count" = "s"."post_count"
            , "post_score" = "s"."post_score"
            , "comment_count" = "s"."comment_count"
            , "comment_score" = "s"."comment_score";

As with all scripts that work directly on the PostgreSQL database: warranty void when connected, your kilometrage may vary, do not look into laser with remaining eye, etc.

Edit: don't count self-votes.