Which ephemeral anchor script should lightning use?

The discussion revolves around optimizing fee strategies in the context of commitment transactions within a specific operational framework. The conversation opens with the suggestion that adopting a P2A approach when the anchor output is of zero value could be beneficial, as it presents a cost-effective method for increasing fees without compromising value. This strategy hinges on the premise that when there's no significant value to be lost (i.e., when the output is considered 'dust'), shifting to a more economical fee strategy like P2A is advantageous.

However, the discourse evolves to address scenarios where the anchor amount surpasses the threshold of being considered negligible or 'dust.' In such instances, the proposition shifts towards integrating the surplus value back into the commitment transaction fee. This approach, although not universally acclaimed, aims at ensuring that the commitment remains zero-fee to allow for an equitable distribution of the channel's assets among its outputs, thereby preventing substantial losses in the form of mining fees during force-close situations.

A critical point of contention within this narrative is the handling of dust HTLCs (Hashed Time-Locked Contracts). One perspective advocates for the preservation of these HTLCs by allocating them only partially to mining fees. This method, particularly highlighted in the second option discussed, introduces a strategic component where peers might outbid each other for commitment confirmation. Such a scenario implicitly requires the winning peer to endure a to_self_delay, emphasizing the tactical dimensions introduced by different fee strategies. This nuanced analysis underscores the complexities and trade-offs inherent in selecting the most appropriate fee strategy, reflecting a broader dialogue on optimizing economic efficiency and strategic positioning within the operational framework under discussion.