(After Bernard of Chartres in the 12th century)
But the point he was making was that his scientific contributions/discoveries were built on what others have done earlier. Yes, it may be necessary to revisit the earlier research (which in the light of later knowledge or observation, may have been flawed) but not always necessary.
In computer science, for example, the earliest compilers were designed and hand coded from scratch (or at least using assembler language). Now, compilers are themselves written in high level language before compilation, and assembler is little used. That said, for really efficient or real time applications, hand crafted code written in assembler may still be more efficient, (drivers are one application that comes to mind) but not many computer users are going to write their own drivers.
Of course, one could argue that the use of poorly written compilers or sloppy coding is contributing to the abundance of bloatware!
One might also argue that it contributes to lack of programmer skills, for example relying (say) on built in referential integrity in a database, rather than coding it in from scratch. Then again, why re-invent a wheel?