This is having into account that:
* Mediawiki is currently serving almost all of the uncached text/api requests
* on 3-5 year old servers
* and not having precached all the pages
* The main purpose of restbase is serving "low latency" simple requests, specially for parsoid and flow
Full results:
https://docs.google.com/spreadsheets/d/1Gtw_i2ZzzuGreElNNRxyxExv4G25sEkonLle7Kj_8ZU/edit?usp=sharing
run:
```
run1:
for i in `seq 1 100`; do line=$(( ( `od -A n -t d -N 4 /dev/urandom` + 2147483648 ) % 12105173)); title=`tail -n+$line /tmp/.fr-nmoWUj/enwiki-20151201-all-titles-in-ns0 | head -n1`; echo $title; TIME1=`date +%s.%N`; curl -s "https://en.wikipedia.org/w/api.php?action=parse&prop=text&page=$title&format=json" | wc -c; TIME2=`date +%s.%N`; echo `echo "$TIME2 - $TIME1" | bc -l`; TIME1=`date +%s.%N`; curl -s "https://rest.wikimedia.org/en.wikipedia.org/v1/page/html/$title" | wc -c; TIME2=`date +%s.%N`; echo `echo "$TIME2 - $TIME1" | bc -l`; sleep 1; done > ~/workbook/comparison1.txt
run2:
jynus@sangai:~/mysql-server$ for i in `seq 1 100`; do line=$(( ( `od -A n -t d -N 4 /dev/urandom` + 2147483648 ) % 12105173)); title=`tail -n+$line /tmp/.fr-nmoWUj/enwiki-20151201-all-titles-in-ns0 | head -n1`; echo $title; TIME1=`date +%s.%N`; curl -s "https://rest.wikimedia.org/en.wikipedia.org/v1/page/html/$title" | wc -c; TIME2=`date +%s.%N`; echo `echo "$TIME2 - $TIME1" | bc -l`; TIME1=`date +%s.%N`; curl -s "https://en.wikipedia.org/w/api.php?action=parse&prop=text&page=$title&format=json" | wc -c; TIME2=`date +%s.%N`; echo `echo "$TIME2 - $TIME1" | bc -l`; sleep 1; done > ~/workbook/comparison2.txt
```