Folding Forum

Looking at this issue report => https://github.com/FoldingAtHome/fah-issues/issues/1237
It seems the issue is fixed in v7.5.2 almost 2 years ago. Where is it and why are we still on 7.5.1?

The same programmer writes the Windows, Mac, and Linux clients, which he has to keep coordinated with the assignment servers, Work servers, collection servers and stats server, all of which he also writes. He seems to be the web designer as well.

He is paid by biochemists, so his focus is on issues important to biochemists. We are not paying him, our issues are well down the priority chain. (Don't you focus first on issues the person paying you wants resolved?)

Do I know what is holding up the next release? Nope, he sure does not report to me, but in my career, I was pushing to get new code out, and the bosses were trying to wait for a graceful time to have errors to troubleshot.

For car dealers, never release code in September, for retail never release code in December, etc. I have no guess when F@H's busy season is. (Other than this last 3 weeks)

JimboPalmer wrote:The same programmer writes the Windows, Mac, and Linux clients, which he has to keep coordinated with the assignment servers, Work servers, collection servers and stats server, all of which he also writes. He seems to be the web designer as well.

He is paid by biochemists, so his focus is on issues important to biochemists. We are not paying him, our issues are well down the priority chain. (Don't you focus first on issues the person paying you wants resolved?)

Hmm okay, fair enough, I understand it's just 1 developer, but I would think 2 years is plenty of time, especially for a minor update? This isn't intended to be criticism, just an observation.

JimboPalmer wrote:Do I know what is holding up the next release? Nope, he sure does not report to me, but in my career, I was pushing to get new code out, and the bosses were trying to wait for a graceful time to have errors to troubleshot.

For car dealers, never release code in September, for retail never release code in December, etc. I have no guess when F@H's busy season is. (Other than this last 3 weeks)

I thought it was "never deploy on Fridays"?

iceman1992 wrote:I thought it was "never deploy on Fridays"?

That is the first rule, followed by never deploy just before a holiday. Everything else are just adds to the first rules.

What I do know, is that features were needed for the servers to support new functions. In addition, about 2 years ago was when F@h went from being based out of Dr. Pande's lab at Stanford with him the Director, to currently being the F@h Consortium with most researchers located at WUSTL, Temple, and MSKCC.

So, much software and servers had to move from Stanford based computers to elsewhere and in the cloud. Those were the priorities given to the developer. Much of that has happened, redoing the server setup to be more "cloud friendly" was next on the agenda as I understand it.

Joe_H wrote: What I do know, is that features were needed for the servers to support new functions. In addition, about 2 years ago was when F@h went from being based out of Dr. Pande's lab at Stanford with him the Director, to currently being the F@h Consortium with most researchers located at WUSTL, Temple, and MSKCC.

I actually missed the story, why the change?

Joe_H wrote:That is the first rule, followed by never deploy just before a holiday. Everything else are just adds to the first rules.

I'm a rule breaker then, sometimes deploying on fridays myself

Accountants seem to have issues with deployments just prior to year end for some reason - honestly I can't see what the problem might be

iceman1992 wrote: Hmm okay, fair enough, I understand it's just 1 developer, but I would think 2 years is plenty of time, especially for a minor update? This isn't intended to be criticism, just an observation.

My largest application (300 users) was all withing a single company, I could get two releases a year out. My brother's largest application, was only about twice as big, but each dealer was an independent company, as was each distributor. He was held up as long as 4 years getting releases out.
In each case, we wrote 3 tier Client server. PC talk to a minicomputer that batches requests to send to a Mainframe, then un-batches the results to send to each dealer's PC.
In my case, one company owned the PCs, Mini, and Mainframe. His client only owned the mainframe.

F@H deals with multiple Universities, each with its own IT Department.

JimboPalmer wrote:For car dealers, never release code in September, for Retail never release code in December, etc. I have no guess when F@H's busy season is. (Other than this last 3 weeks)

iceman1992 wrote:I thought it was "never deploy on Fridays"?

That was our rule at the gold mine, The mine ran 24/7 and if you deployed on Friday, 6 shifts of miners hated you. Always deploy near the beginning of a 12 hour shift. That way only 500 employees suffer.
https://www.barrick.com/English/operati ... fault.aspx

iceman1992 wrote:I actually missed the story, why the change?

Dr Pande took a private sector job about 3(?) years ago, but maintained his lab at Stanford for a period of time until his grad students and post-docs finished up. He and Dr. Bowman worked together for a period of time to transition to Dr Bowman heading up the F@h Consortium. There are still a few people at Stanford still active with F@h, but as far as I understand it, no active research lab.

There are some posts on this if you scroll back far enough in the News blog. There may also have been some posts here, it gets a bit hazy from the "too many sources" of information I follow.