For the lack of logic, just think about how your own mining client works. Most of you readers (yes, not all, but most), are mining at a traditional pool. You consume very little bandwidth, you dont need to store the blockchain and you dont need to know or verify any TXs. Yet your mining effort works towards blocks with transactions included.
You work successfully under exactly the same conditions, as (some of) you claim would force a botnet to exclude transactions. (Some of) you say, that downloading the blockchain would make botnet victims suspicious, yet you do not need to download it yourself! (Some of) you say, the steady influx of transactions would need lots of extra CPU processing and extra network bandwidth, yet you do not need to do that yourself! (Some of) you even say that a full bitcoind must be in place, yet you do not need it on your own miners!
Your own mining client shows how the technical "problems" can be solved, and source code is available for everyhing (both mining client and pool service).
The assumption of the technical benefit is just plain false. The only thing that supports the idea of a botnet, is the ever repeating posts here on bitcointalk. Those who jump to early conclusions and keep reinstating this false idea, are the ones who do damage to bitcoin.
In a dumb miner - pool model the fact that the miner doesn't do the workload doesn't mean the workload doesn't exist.
One getwork is ~400 bytes. Some have speculated mystery has 1.8 million nodes. To use a traditional full pool server the bandwidth requirements alone would be 5.7Gb and 1.8 million connects every long poll. Say he staggered those getworks to supply all nodes with work over a 10 second period (less revenue btw). You are still talking about almost 200K simultaneous connections, 750Mbps, and enough computing power to generate 200K block headers a second.
If you want more insight ask Tycho what kind of load even small botnets do to a pool server. Now with enough computing power (we are talking a dozen or so high end load balanced servers on a 1 Gbps connection) a 1.8 million node botnet "could" operate in a pool-miner model. However now compare all that cost vs the incremental reward of .... ready for it ....
about 2 cents per block.