Okay, okay. I can hear the cry of thousands of Rosetta crunchers calling to me.

As suggested earlier by feet1st, This thread here will draft the first of three letters. The second and third letters will be drafted in another separate thread for each.

The letter project is something I think I can help with, but I will need the help of others, especially since my typing skills are measured in integers. As I see it from the conversation so far with my idea, there are several things that a serious email should cover regarding inactive users. They are put here in no specific order:

1. The bug issue has been fixed and we need your help, please come back

2. The project has had many successes and the best is yet to come

3. The project has tangible benefits for disease research

4. Consider giving Rosetta higher priority because of the benefits

If a volunteer can draft any of these paragraphs, post your idea here and we will all get to see the letter as it develops. The opening paragraph needs to be short and pithy for those who just want to skim the letter, and as usual introduce all of the above issues as well as summarize them for quick reading.



On another matter, I haven't followed other threads too much, but has the issue of high memory requirements been addressed? I was thinking that the server could be programmed to distinguish between hosts, so those with low memory of less than 512mb of ram could be only given those WUs that don't hog the ram so much. Just a thought.


As to the size of the data units, I haven't seen anything on this matter either. Please advise as to the status of this issue.

I like your outline. I would add:

"Thank you for your participation with Rosetta@home in the past."

...has the issue of high memory requirements been addressed?

There are two threads in the crunching forum that touch the topic, so let's take that off this thread and focus on the letter. Here are links for you:
First thread
Second thread

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/

i would wait a while (say >=1 week) just to ensure that the bugs (esp. 1% bug) have well and truly has been squashed. you don't want to say we've fixed the bugs so please come back, but they haven't been fixed properly. if this happens, you probably will never, ever get them back to rosetta.

---

i would wait a while (say >=1 week) just to ensure that the bugs (esp. 1% bug) have well and truly has been squashed. you don't want to say we've fixed the bugs so please come back, but they haven't been fixed properly. if this happens, you probably will never, ever get them back to rosetta.

You are right.

But why not make it ready and then send it at the right time. :)

Anders n

---

But why not make it ready and then send it at the right time.

Exactly. If ya don't got it, ya can't send it. It'll take a week just to get feedback and make revisions.

A few other things I suggest on the outline

Have you seen Dr. Baker's journal with science udpates?

Rosetta needs more computing power because there are new projects coming online to study HIV and malaria in addition to the base protein folding work they have been doing.

"If you have any doubts or questions, please review the FAQs or post on the message boards, we'll be glad to help you crunch more Rosetta."

"Please help a friend get started on Rosetta too." (would be great if we had the "tell-a-friend" function implemented and could just link to it.)

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/

We've lost some due to the bandwidth issues, and obvious problems such as the 1% bug.

On the bandwidth usage front, I was seeing 50-100 megs/day when I first started. Now we can limit our systems to running just one WU per day, and get closer to 5 megs/day. Which would mean 150 megs/month. Soon we'll be getting a new Boinc client that should allow the Rosetta client to shrink the bandwidth usage almost in half. (And if Boinc doesn't get that out soon enough, all our major bugs will killed by Rom and the rest of the Rosetta team, and we'll get better compression added directly to Rosetta.) If you left because of bandwidth usage, please return and stop by to ask for help setting your system up to have a max cpu time of 24 hours.

As of March 31, 2006 (for International users, it's easiest to spell the month out), we've lost around 60% of the errors returned to the Rosetta server. It dropped from around 10% down to around 4% for the Windows clients. Testing on the hanging at 1% bug continues, but there's good news: One group of 40 systems was usually having 4 systems reported with WUs stuck at 1% each inspection. Since the release of the Rosetta 4.83 windows client, those 40 systems have gone 36 hours without a 1% stuck WU being caught. Even if you do end up with a WU stuck at a low percentage, changes to the program will now show where the error occured in much more detail (3.1415%, with 3 or 4 digits after the decimal point). Rosetta will now kill WUs that don't progress far in 24 hours, so the most you can lose now is 24 hours of runtime. If the hung WU bug isn't found soon, it's been promised that they'll bring in a dominatrix from Rachael's Dungeon of Pain, and her 8 foot long whip should encourage Dom to track down the bug even faster*. (*Okay.. so maybe they haven't. Publicly. :)

Some of those that have left - will have left for reasons we haven't touched on. Ask them to stop by and see if the things have changed so they feel like contributing again. And for the competitive amongst them, let them know we'll need their help during Casp 7 to deal with some of the larger/more work intensive proteins where we need to process 10 times the models we currently do for a useful result.

And for those that like to suggest improvements - encourage them to return and do so. We've managed to get a few of our suggestions implemented in the Rosetta client so far.

---

I was hoping that someone would post a paragraph on one of the issues the letter needs to deals with. I will post the introductory paragraph Wednesday if at all possible; perhaps a few are waiting for me to take the lead here. Trying to get through a semester that is difficult is slowing my efforts here down substantially, so please be patient.

If I can I will do so then. In the meantime, those who want to contribute please get to dealing with one of the issues mentioned below in a paragraph. To be sure, it would be difficult for me to imagine that I can write paragraphs for some of the issues needed, so I definitely need you help.

Gerry

gerry,

feets and i have worked on a "tell a friend" email that we hope will be approved for the email link. we've started to tackle this one as well, gathering information as to the best way to approach it. give us till the weekend and then we'll post what we've come up with, or if feets is uncomfortable with that he'll chime in here.

---

I was hoping that someone would post a paragraph on one of the issues the letter needs to deals with.

Gerry, welcome to the Rosetta volunteers! ...As you can see, it is sometimes difficult to get simple things done. But, one at a time, they ARE getting done!

The volunteers group vavega refers to is presently focused on completion of the tell-a-friend function, which includes the text of an EMail to send to someone that knows nothing about Rosetta.

Please Gerry, and those reading this, come share your opinions on the work we're doing. We'd love to work through the glitches and present truely complete work back to the project team, but presently don't have enough voices to do so.

We're hopeful that once that "tell-a-friend" letter is finalized, it can morph in to: A) a letter to inactive users, and B) a letter or a paragraph that we will request other projects include, as their work winds down for a period of time, in a note to their participants.

Due to the timeline on Einstein, I'm thinking we might best do "B" first. They've only got about 5 or 6 weeks of WU left, and they continue to add more users. Also, it will take a couple weeks to try and persuade them to forward the Rosetta msg with their own "thanks, end of work units, come back later..." type of msg... if they're even planning to send one.

Einstein is presently more than double the TFLOPS of Rosetta, so there are more users there then ever were over here. Also that might afford more time to assure the nail is set on the 1% bug. Or is that 1.042%? To deliver a more confident msg to the inactive users.

If you haven't seen it, here is one of the many "hidden" links on the Rosetta pages. We can't offer ice cream, (which MAY be just as well, because I don't think the phrase "BOINC for ice cream" would work outside of the college campus for which it was intended) but it's an example of someone else's effort to convey the compelling msg.

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/

Gerry: Here's an attempt at what you were asking for - and should at least be a starting point to help bring in critiquing, improvements, and more detail. (And inspire others to join in.)
----------------

Rosetta@Home wants you back. The most aggrevating bugs have been eliminated, and we can use the extra help. Rosetta did good at the CASP 6 competition in 2004, and with all the improvements that have been made, even in the last few months, we should do much better with this year's CASP 7 competition. As you know, one of the purposes of Rosetta@Home is to improve the Rosetta client so that it can be used to help find treatments and cures for various diseases and ailments that we humans suffer from. The Baker Lab is working on projects like a vaccine for HIV. DC projects are using the Rosetta client to help target other diseases and ailments. Consider giving Rosetta a share of your Boinc cpu time, or returning to Boinc so your aid can help bring these goals to fruition much faster.

On the bandwidth usage front, Rosetta was using 50-100 megs/day earlier this year. Now we can limit our systems to running just one WU per day, and get closer to 5 megs/day, which would mean 150 megs/month. Soon we'll be getting a new Boinc client that should allow the Rosetta client to shrink the bandwidth usage almost in half. (And if Boinc doesn't get that out soon enough, all our major bugs will killed by Rom and the rest of the Rosetta team, and we'll get better compression added directly to Rosetta.) If you left because of bandwidth usage, please return and stop by to ask for help setting your system up to have a max cpu time of 24 hours.

As of March 31, 2006, we've lost around 60% of the errors returned to the Rosetta server. It dropped from around 10% down to around 4% for the Windows clients. Testing on the hanging at 1% bug continues, but there's good news: One group of 40 systems was usually having 4 systems reported with WUs stuck at 1% each inspection. Since the release of the Rosetta 4.83 windows client, those 40 systems have gone 36 hours without a 1% stuck WU being caught. Even if you do end up with a WU stuck at a low percentage, changes to the program will now show where the error occured in much more detail (1.042% for example). Rosetta will now kill WUs that don't progress far in 24 hours, so the most you can lose now is 24 hours of runtime. If the hung WU bug isn't found soon, it's been promised that they'll bring in a dominatrix from Mistress Rachael's Dungeon of Pain, and her 8 foot long whip should encourage Rom to track down the bug even faster*. (*Okay.. so maybe they haven't. Publicly. :)

2.
Please stop by and see how the client has improved. For the competitive amongst you, we'll need your help during Casp 7 to deal with some of the larger/more work intensive proteins where we need to process 10 times the models we currently do for a useful result. With several of the most recent changes to the client, we're guaranteed to do much better in CASP 7, however, with enough cpu power we may be able to come in 1st place for all the proteins we fold this year. By processing a protein once to learn the energy landscape, we'll be able to process it a second time in high detail mode with the speed of the low detail mode. This should improve our protein predictions greatly. With proper motivation, the Rosetta team will continue dreaming up new approaches, testing them out with your processing help, and incorporating the ones that improve the results back into the Rosetta client.

3.
As the Rosetta client improves, and the predicted results come closer and closer to the native structure of the protein, the client will be used in more DC projects to help identify cures and treatments for various human diseases and ailments. At present, the Rosetta client is being used in [World Community Grid?] as well as {fill in blank}. Each improvement in the code brings us closer to the goal of a client that can be used to rapidly identify potential treatments for various ailments that can be used by the medical industry. Perhaps even finding a cure of over-wordedness and technobabbling. /e ducks.

4.
In summation: Please consider returning to Rosetta and helping benefit humanity. We need you, the Rosetta client needs you, and those that will benefit from the future use of this client need you. The life you help improve may well be your own.

----------------

I'm starving to death.. and need to get out of here.. so if my humour circuits are failing and I seem way too singsongy and melodramatic.. I blame it on dreaming of dinner.. :)
Somebody that's not starving and able to concentrate should be able to do a better job.. :)

---

I can hear those voices in my head again, telling me that Rosetta crunchers are calling me to serve their cause, sacrificing my body in the service of... Well you get the idea ;) In any case, I've now joined the rosetta volunteers who are trying to get the letters project(s) rolling. I recommend that viewers of this thread join that group as well and help out in any way you can. It only takes a few minutes: click the volunteers list link below and follow the instructions.

As to the submission immediately below, I like much of it; I would submit it for discussion on the discussion list. Let's see what is out there and decide this together.

In the meantime though, it is best to close down this thread and get things over on the discussion list where there are others waiting to here from the rest of us. See you on the list!!

Gerry Rough

One suggestion I'd like to make to people working on a letter, is to use the simplest language possible. Remember for many English isn't their native language. Also, keep it short.

There are many studies which show that people basically "skim" over the text.

I think you'll find a wealth of information in the following report:

http://www.useit.com/alertbox/20040217.html
By: Nielsen Usability Studies

"Summary:
E-newsletters that are informative, convenient, and timely are often preferred over other media. However, a new study found that only 11% of newsletters were read thoroughly, so layout and content scannability are paramount."

PS: He has some good tips about Web design. Personally I find Rosetta's Website excellent, easy to navigate (with the main links on top at all times) and aesthetically pleasing.

Some issues which IMHO still need to be addressed are: simpler 1-2-3 install instructions with Project URL readily available, and some "consolidation" and "index" of science material, some of which takes a bit of looking around.

---

Best UFO Resources
Wikipedia R@h
How-To: Join Distributed Computing projects that benefit humanity

with Project URL readily available,

I take it that means you are in the same boat that I was in for weeks. I hadn't noticed that the heading on the homepage ABOVE the rules and policies takes you right to info with the project URL here. The only other link I've found that takes you there is the one at the end of the introductory paragraph where it says "please join us". But, I'm not sure I'm READY for that link yet at that point... I'd prefer one to more science info as you suggested when you said:

and some "consolidation" and "index" of science material, some of which takes a bit of looking around.

At that point in my absorbtion of the project, I'd like the laymen's guide to what you are doing, what you hope to achieve, and why you need my help with that. And at the end of THAT page, a link to get in to the "how BOINC works" stuff, "why are you sending me to some OTHER website to download software?" and please register for a Rosetta account, which leads to "why do you need this information?"

As you can see in this thread, it takes more than a week just to write a reasonable letter. The easy CD is on the todo list, but certainly requires much more work than the letters if we're really going to do it up right.

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/

on the homepage ABOVE the rules and policies takes you right to info with the project URL...

I had noticed it, but if I were designing a BOINC project site, I'd try to keep in mind that even though the word URL (=Uniform Resource Locator) maybe clear to ME (myself being on the Internet before the Web was even available and using NCSA Mosaic and Netscape 0.9 and 0.x versions of Linux etc) even the word URL might not be clear to "everyday" people, who are using the word "web address" instead.

Let alone the mixups which might occur due to the fact that the Rosetta@home URL includes a path "/rosetta" vs just the base domain a la RALPH (ralph.bakerlab.org) etc.

I have some experience in these issues (one of my ventures had to do with medium sized e-commerce sites and we did usability testing for many different types of navigation etc) based on that experience my motto is KISS = Keep It Simple, Stupid, 1-2-3 type stuff ("guiding" you through the process e.g. the shopping carts of Amazon)

As I said, IMO Rosetta's website is already very nice, but could use a little bit little of cleaning up. And, personally, I find the Einstein howto-join layout more intuitive.

PS: I'll try to improve Rosetta's Wikipedia page
http://en.wikipedia.org/wiki/Rosetta%40home
this weekend (incl. some Q&As from journal) so every feel free to offer feedback before I start re-arranging things.

---

Best UFO Resources
Wikipedia R@h
How-To: Join Distributed Computing projects that benefit humanity

I had noticed it, but if I were designing a BOINC project site, I'd try to keep in mind that even though the word URL (=Uniform Resource Locator) maybe clear to ME (myself being on the Internet before the Web was even available and using NCSA Mosaic and Netscape 0.9 and 0.x versions of Linux etc) even the word URL might not be clear to "everyday" people, who are using the word "web address" instead.

I've already started looking into this. We've asked the UWashington about using "rosettaathome.washington.edu" - but things take time. rosettaathome.bakerlab.org already works... this all seems a bit long to me as well. Perhaps folks would like suggest ideas...?

---

rosetta.bakerlab.org would sound just right to me... but then maybe you were thinking of something even more simple? Like www.rosettaathome.org ?

Just my 2c

- Trib

---

rosetta.bakerlab.org would sound just right to me... but then maybe you were thinking of something even more simple? Like www.rosettaathome.org ?

Just my 2c

- Trib

Edit: Sorry for the double posting here. =(

---

rosetta.bakerlab.org would sound just right to me...

I agree, either

rosetta.bakerlab.org (would be my first choice) or
rah.bakerlab.org

short, easily spell-able etc. (or .washington.edu)

---

Best UFO Resources
Wikipedia R@h
How-To: Join Distributed Computing projects that benefit humanity

given it just has to redirect (I know leiden classical uses a redirect as well)

boinc.rosetta.org
or
rosetta.boinc.org

would be my vote if it was viable,
next would be (and I beleive should be/will be ;)) whats used

rosetta.bakerlab.org


(why my second choice ? I keep typing bakerlabs :-s )
why I beleive it should be it anyway.
- it goes well with ralph.bakerlab.org
- bakerlab get some advertisment
- it is simple (if you can spell)
- it looks neat
- bakerlab.org domain is already owned

---

Team mauisun.org

Perhaps folks would like suggest ideas...? (for project URL)

This subject is an important question, but off topic. If the project is going to change the URL, please let me know before I buy 10,000 bumper stickers that promote a URL that's been changed. Just my $2k worth (the approximate cost of that many bumper stickers).

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/