On 01/10/2014 11:25 AM, Kian-Tat Lim wrote: > The position I'd like to take is that we will figure out a way to make > LSST Stack packages easily installable outside the Stack-standard > environment. This I would love. I'm really just trying to simplify development, and I see making stack packages easily installable as a task I'd rather not pick up, largely because the packaging system is still opaque to me. > An alternative is that a "yum install" of qserv actually installs the > Stack-standard environment, and that qserv lives within it. Initially > this seems less desirable, but it may not be if the environment > installation is made lightweight enough. Again, this is probably fine, but packaging up the stack environment into an rpm/deb feels kind of hard to me. But it might not be. It could be like installing pick-your-scripting-language's own package manager (which sysadmins sometimes accept, even as it annoys them). > Having qserv go its own way to the extent that it cannot easily share > code with the LSST Stack is simply unacceptable. Philosophically, I agree. Even if LSST will perpetually use eups-managed LSST stack code, what you propose above (making packages easily installable outside the stack-std env) is immensely valuable. A couple of possibilities: a python-lal (LSST Astronomical Library) package or a liblsstr (lsst runtime library, like libapr). This would make it easier for an open-source planetarium package that wants to reuse some algorithms. If possible, I think it's best if we have an eye to keeping things generic so our software plays well with everyone else, as long as being generic doesn't cost us too much in complexity. Anyway, I have the (probably mistaken) impression that in order to reuse stack code, the normal way is that application X becomes a stack package, at least on the local machine. I'm sure the code is reusable more easily than that. I can write up some doc for how to do this and put it in a wiki page, as long as someone can give me the highlights and work with me. Sound okay? -Daniel ######################################################################## Use REPLY-ALL to reply to list To unsubscribe from the QSERV-L list, click the following link: https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=QSERV-L&A=1