The original question was to predict the basic requirements for information storage. Then the discussion moved to why is it necessary to include mRNA in the protein translation process. Why can't there be a similar chemical molecule like tRNA which can directly read from DNA and exclude the mRNA intermediate. Also is the current translation process optimal in the sense that can there be a system which does not need mRNA or a similar intermediate altogether. Or is the presence of an intermediate a necessary requirement.
One reason is that an intermediate like mRNA allows for higher amounts of protein expression. You can have multiple mRNA molecules that are translated simultaneously. If you read directly from DNA you can have at most two translations in parallel.
I'm not sure about this, but I would imagine that having to unwind the DNA double strand every time for translation could be a drawback to directly reading from DNA as well. The mRNA is also subject to various forms of regulation, which might be an advantage as well for the added intermediary.
I don't have a lot of references for this, but it's too long for a comment.
Separating the roles of RNA and DNA helps to better control protein production and gene replication. If ribosomes worked directly on DNA, it would probably be very hard to replicate that DNA, as the DNA polymerases would collide with the ribosomes. You'd have to stop protein production and move the ribosomes out of the way before you could copy the DNA. Keeping mRNA separate allows the ribosomes to keep making protein while the DNA is being replicated.
However, transcription of DNA would be similar, where the DNA replication could collide with the transcription machinery. I can't find anything about that situation so I don't know how its regulated or how often it happens.
Additionally, having mRNA allows amplification and further regulation. Every mRNA copy can make several protein copies, so the amount of protein produced from a single round of transcription can be higher than if the DNA were read directly by a ribosome. mRNA also has many binding sites for miRNA and other regulatory RNAs. If DNA were read directly, that wouldn't be possible.
There are also some cases where mRNA is stored and used later, such as immediately after fertilization, or in platelets and erythrocytes where translation can happen even after transcription stops due to loss of the nucleus.
In lay man terms you need your DNA to remain intact and protected. so instead of shipping it to the cytoplasm where a lot can happen to it, the cell makes an intermediate mRNA. if anything does wrong with mRNA, the cell can discard it and make another copy and after translation the copy of mRNA is also discarded.You also have the amplification; instead of making one protein from one gene on a DNA strand, and then one more and the another, we can just make a bunch of copies of mRNA and send them to the nucleus where we can make as many proteins as we need. by doing this there is minimal manipulation of the original DNA strand and the cell get a lot of product over a short period.