In a way this is like LTC being the Debian "Stable" and Bitcoin more of a Testing/SID; I think we should not move to far up the stream, security fixes excepted.
I totally agree. If there's interest from other miners we can try maintaining our own "people's Litecoin" client based on 0.6.x. Most of the problems cited above should be solvable in other ways (like you suggested, we could backport fixes if there are particular issues to be concerned about).
I'm concerned about the Litecoin development right now. If we start pulling in Bitcoin changes the Litecoin community will have no say on the direction in which the client develops. We'd be completely dependent on the Bitcoin guys at that point. In addition, while I appreciate the efforts put in by the Litecoin Dev team, it's not a good sign if all of the new Litecoin development happens behind closed doors.
Are any other miner's interested in a separate 0.6.x codebase that continues the spirit of the original client? If so we could start a 0.6.x branch that fixes the most pressing issues. The "15th May" issue discussed above could be easily solved by miners putting a hard-limit of 250k on the blocksize. This would allow a more gradual and proper fix, without any risky hard-forking.
You are expressing some common misconceptions.
The Bitcoin hardfork on 15 May was to prevent the type of accident that occurred on March 12th. The post-mortem analysis of that accident however revealed an entirely different vulnerability that under certain unlikely conditions would cause identical clients to permanently lose consensus and fork from each other in a failure of self-consistency. Satoshi didn't realize BDB lock limit, so assumptions everyone had long held about the behavior of the protocol turned out to be incorrect.
Prior to the May 15th deadline, Bitcoin-0.8.1 imposed two temporary rules. First, it rejected blocks with enough tx's that would trigger the difference in consensus between BDB and LevelDB clients that was the March 12th accidental fork. Second, it reduced block size generation limit to 250KB. The temporary 250KB soft limit of Bitcoin block size prior to May 15th reduced the likelihood of the unlikely forking condition to occur, but it did not eliminate the risk. The protocol allow blocks up to 1MB in size, a determined rogue miner with sufficient hash firepower could artificially trigger it. (I am not describing exactly how to do this attack in detail. Suffice to say it is exceedingly difficult to pull off, although not anywhere near 51% difficult.)
So no, your suggestion of 250KB block size limit wouldn't fix the problem. If you meant to modify the protocol to impose a permanent hard-limit of 250KB such that larger blocks would be rejected, that would be a hard fork from all current Litecoin-0.6.x clients. Didn't you mean to avoid hard forks?
Bitcoin would have strongly preferred to avoid a hardfork if they could. The surprise of the BDB lock limit really messed up their earlier plans to give people 2 years of warning before protocol hardforks. The fact that Bitcoin considered it mandatory to go through a hardfork so soon suggests the severity of this issue.Alternative: Backport Litecoin-0.6.x
Bitcoin issued backported patched builds of versions 0.4 through 0.7 containing the same lock limits and temporary rules until May 15th as Bitcoin 0.8.1. We could similarly do the same and issue an official Litecoin-0.6.4 that will survive the hardfork. The trouble with calling the backport "official" though means a significant amount of extra work to backport and test to ensure this will remain compatible.
You might have noticed Litecoin has had no releases for a year. Developers are busy with day jobs, and my time working here is very limited. I personally think there is no advantage to using 0.6, so why would I spend my limited time ensuring it works properly? Otherwise it probably shoud be released as a "best effort" git tree with a "use at your own risk" label, which is pretty much what any unofficial client fork should be warning anyway.Stop Blindly Worrying about Pulling Code from Bitcoin
Where is this crazy talk coming from? Litecoin is a copy of Bitcoin with some key enhancements (mainly anti-spam). Litecoin can benefit from Bitcoin's constant evolution, where their codebase becomes faster, more efficient, and they fix bugs. Litecoin does not blindly pull all code from Bitcoin. This development is a very time consuming process because I personally walked through all historical Litecoin patches, and all Bitcoin patches since 0.6.x, thinking in each case "Is this good for Litecoin?" and "Does this work together?" Some parts of Bitcoin-0.8.x will be disabled or changed in Litecoin. The internal git tree currently is on its 4th rebase in the last two weeks as I better understand the codebase, clean up past mistakes and split Litecoin diffs into more easy to understand functional groupings to allow for better future code maintainability.
About the "behind closed doors" concern. That is fair concern. Currently we prefer to keep the tree hidden because it is experimental and quite dangerous for users who will try to use it and screw themselves over. When it is cleaned up and in better quality, there will be a public beta period of 1-2 months prior to an actual release during which the public is encouraged to test and comment on the code.Vote with Your Wallet
If you appreciate my approach, please donate. Identify yourself in PM if you want your name to be on a list of donors. The list of donors will need to be published much later in order to obscure the identity of donor addresses. If folks donate a meaningful amount and identify themselves with a crypto signature of the sending address, I am more likely to consider the merit of their ideas. If the idea is bad, I will at least respond with why I believe the idea is not workable.
If you think my ideas are bad, then feel free to launch your own competing project to become the next Litecoin reference client. coblee will have to approve whatever is proposed, whether it is my git tree or your alternative. If you disagree with coblee, you are free to try to convince users to follow your chain fork instead of the litecoin.org chain.