Youtube comments of Maksym Barbul (@maxbarbul).
-
445
-
299
-
27
-
24
-
19
-
13
-
10
-
9
-
5
-
5
-
4
-
4
-
4
-
4
-
4
-
4
-
3
-
3
-
3
-
2
-
2
-
Thank you for the back story, it was entertaining to watch. I believe, progress bar is mostly about perception and ux than accurate estimate. It still feels like it is possible to make more or less stable prediction after copying has been prepared and analyzed. You should know by that time the spread of file sizes and disks being involved. So, you can adjust estimate based on known overhead. If some heavy disk load comes at the time of copy than it’s a unexpected circumstance (and user might be aware of it or made aware of).
Comparing to driving, despite roads you’re going by have limit like 55, 35 miles/h, etc, there is an average speed which is around 25 miles/h so, directions algorithm should assume you’re going this average speed.
Anyway, thank you for explaining all the details involved, it is in deed a hard task.
2
-
2
-
2
-
2
-
2
-
2
-
2
-
2
-
@luckyluke1503 hm, how did it happen that russia was established in 1991 and nukes before that belonged to this not existing country?
There was one dictatorship USSR, there were Rssr (russian soviet socialist republic), ussr (ukrainian soviet socialist republic) and other 13 republics, including moldovian, kazah, turmen, and many more. Those republics were one “unbreakable” country with same laws, language, central government. Russian language was the state language because, the revolution started in russia and after that they occupied other states, they damaged cultures of many countries.
Go learn history
2
-
2
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
Great video, however i would say some points in cons and pro’s could be softened.
I.e., you can scale mysql horizontally with replicas; you have to think about schema and migration in both paradigms as everyone needs robust and consistent applications. You need to write migrations for nosql and support all variations of data in your app, this may be a not easy task.
Sql data normalisation wasn’t invented to save data, rather to avoid data anomalies. You can have denormalised tables for optimisation purposes. Indexes allow to optimize read speed in sql. Etc.
Also, while comparing, data amount matters. Are we building whatsapp? Then we need every nanosecond we can get. If it’s a regular blog or ecommerce site with tens of thousands items, you don’t need sharding. You’d spend more time designing partition key and indexes for nosql then in sql like posrgres.
If you have a lot of log-like data, in sql to speed up writes, you can rotate the table like logs, have many tabels with suffixes cut by date or something.
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
Great video, however i would say some points in cons and pro’s could be softened.
I.e., you can scale mysql horizontally with replicas; you have to think about schema and migration in both paradigms as everyone needs robust and consistent applications. You need to write migrations for nosql and support all variations of data in your app, this may be a not easy task.
Sql data normalisation wasn’t invented to save data, rather to avoid data anomalies. You can have denormalised tables for optimisation purposes. Indexes allow to optimize read speed in sql. Etc.
Also, while comparing, data amount matters. Are we building whatsapp? Then we need every nanosecond we can get. If it’s a regular blog or ecommerce site with tens of thousands items, you don’t need sharding. You’d spend more time designing partition key and indexes for nosql then in sql like posrgres.
If you have a lot of log-like data, in sql to speed up writes, you can rotate the table like logs, have many tabels with suffixes cut by date or something.
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1
-
1