FWIW, cost to run Bitcoin full node is much lower than cost to run node for current top 100 altcoin by market cap.
Bitcoin rocks! I was distracted by altcoins when first finding out about bitcoin, and had to learn over time that bitcoin is still the best for many things.
While it's theoretically possible to propose extreme fork which do these,
1. Treat script OP_FALSE OP_IF ... OP_FALSE as invalid.
2. Add size limit to script size.
3. Treat OP_RETURN with data bigger than X bytes as invalid.
4. Treat new UTXO that contain less than X satoshi as invalid.
Such fork is very unlikely to be accepted. And if it does, some spammer will continue to add arbitrary data using fake address at higher cost.
In my opinion, that's a great feature of bitcoin. It prevents a lot of bad code changes from getting introduced.
Is it possible to detect transactions that contain large UTXOs, large numbers of public keys (some may be fake, and we don’t know which ones are fake), or other attributes that may identify a transaction as harmful?
AFAIK no. Although i recall there were few discussions here that talks about theoretician address format which make fake public key/address become more expensive.
I'll have to look for some of those discussions. It sounds interesting.