This is a Gist for stuffing my quick notes and thoughts in here, also serving as development tracker for bits.
Now bits has moved to WordPress.
Key points:
-
README.md
isn't very important since it isn't a note, but it may be used for some indexing or linking to other years. -
Each note is a comment, currently group year's notes in one Gist. No need to title each note, too much effort for reference when comment link is enough.
Originally, I thought of using pattern like
YYYYnnn
, year with 3-digit number. But it's really not necessary. Not even me would actually refer to the notes. -
Must have proper formatting capability, like Markdown.
Although these are just notes, but I would prefer formatting them well. Not tweets, nor blog posts.
-
Must be able to scroll down a single page to view multiple rendered notes.
I wished to have each note their own page, but couldn't really find any easy solution to meet my needs, so settling down in Gist comments works best for now.
-
Okay to lose these notes, they may not be worth reading again.
-
Okay not to have proper search function since each note is a comment.
-
Okay not to have separate comment thread for each note.
-
Okay not to have commit history for notes.
Ideas:
-
The first note is a bit long due to the emerge log, it might be a good idea to have another Gist to store those log stuff, like
bits.2013.bits
. -
keyword' bookmark for comment form
Unfortunately, the form doesn't have ID.
Thinking about using YouTube API v3 to perform a random video search. Since there is no built-in option for random, I thought I could use that
publishAfter
andpublishBefore
by combining with a special made pdf like by month. Starting from February, 2005, when YouTube was found.But when I entered
2005-02-01T00:00:00Z
intopublishBefore
, there are about 90K results. As I looked into the date, I realized that's not the published date as video uploaded date. It seems to be Recording date which uploader can enter any date they want.Honestly, I don't know who made that decision when on video pages you can read "Published on M, d YYYY." That's really a "published" != "published."
And I am calling that decision is terrible, why? Because they you entered UNIX epoch
1970-01-01T00:00:00Z
you would get total results of 1,000,000. The cap value when there was a lot of video hits.See? You can't use pagination, that next page token is just, well, awful design to say the least. It comes with its partner in crime, "infinite scroll." When infinite scroll showed up on Internet, accuracy and precision were thrown into fire. Next, next, and next.
How many times do I have to remind my self: stay away of APIs?