hi, Thomas, Christian, Dan, could you please help me test the http://ras.openlib.org/ It runs the latest ACIS with some corrections / fixes, and RAS data & details. It does not have much data in terms of papers, but some, and some citations. It does not run APU right now. Known issues: * http://ras.openlib.org/privacy-policy and some other pages do not work, but http://ras.openlib.org/about and /faq work -i
Ivan Kurmanov writes
hi,
Thomas, Christian, Dan, could you please help me test the http://ras.openlib.org/
It runs the latest ACIS with some corrections / fixes, and RAS data & details. It does not have much data in terms of papers,
did the problem with the update daemon crashing reappaer?
Known issues: * http://ras.openlib.org/privacy-policy and some other pages do not work, but http://ras.openlib.org/about and /faq work
Add this provfile-overview 404s. http://ras.openlib.org/profile-overview!e7de25b4 but this again is probably a sysadmi rather than ACIS issue. Cheers, Thomas Krichel http://openlib.org/home/krichel http://authorprofile.org/pkr1 skype: thomaskrichel
On Fri, Nov 4, 2011 at 11:51 PM, Thomas Krichel <krichel@openlib.org> wrote:
Ivan Kurmanov writes
hi, It runs the latest ACIS with some corrections / fixes, and RAS data & details. It does not have much data in terms of papers,
did the problem with the update daemon crashing reappaer?
I've fixed the problems with update daemon crashing.
Known issues: * http://ras.openlib.org/privacy-policy and some other pages do not work, but http://ras.openlib.org/about and /faq work
Add this
provfile-overview 404s.
http://ras.openlib.org/profile-overview!e7de25b4
but this again is probably a sysadmi rather than ACIS issue.
it is, actually. Apache configuration issue. -ivan
In /affiliations we have Search institutions click here the anchor text is meaningless and document rather than subject talk. "more about author affilitiations" would be better Cheers, Thomas Krichel http://openlib.org/home/krichel http://authorprofile.org/pkr1 skype: thomaskrichel
Thanks to all who tested! The updated list of known issues: * http://ras.openlib.org/privacy-policy and some other pages do notwork, but http://ras.openlib.org/about and /faq work * profile-overview page returns 404 error * "click here" link on the affiliations screen * name variations on the research overview screen are shown with toomuch spacing between them Anything else? -ivan On Fri, Nov 4, 2011 at 11:56 PM, Thomas Krichel <krichel@openlib.org> wrote:
In /affiliations
we have
Search institutions
click here
the anchor text is meaningless and document rather than subject talk.
"more about author affilitiations" would be better
Cheers,
Thomas Krichel http://openlib.org/home/krichel http://authorprofile.org/pkr1 skype: thomaskrichel
_______________________________________________ RAS-run mailing list RAS-run@lists.openlib.org http://lists.openlib.org/cgi-bin/mailman/listinfo/ras-run
Christian and all, I've done some more work on ACIS & RAS. I've integrated most template (XSLT) changes that Christian did on RAS, both to screens and emails, into the main (Thomas') ACIS version. None of the RAS-specific changes went into the default template set (presentation/default/). I've created a new phrase.xml file based on the one I took from nebka. Thomas may find these patches in the aras@holda.openlib.org:dev/acis.r repository, just as before. Also, I've fixed the issues listed in the previous mail, see below. To me this version looks as if it is ready to be installed on RAS. Other problems may arise when we do this upgrade, but it is very hard to diagnose them now. So, I suggest to do the last round of testing and then install it. -ivan On Sat, Nov 5, 2011 at 10:32 PM, Ivan Kurmanov <duraley@gmail.com> wrote:
Thanks to all who tested! The updated list of known issues: * http://ras.openlib.org/privacy-policy and some other pages do notwork, but http://ras.openlib.org/about and /faq work * profile-overview page returns 404 error * "click here" link on the affiliations screen * name variations on the research overview screen are shown with toomuch spacing between them
Anything else? -ivan On Fri, Nov 4, 2011 at 11:56 PM, Thomas Krichel <krichel@openlib.org> wrote:
In /affiliations
we have
Search institutions
click here
the anchor text is meaningless and document rather than subject talk.
"more about author affilitiations" would be better
Cheers,
Thomas Krichel http://openlib.org/home/krichel http://authorprofile.org/pkr1 skype: thomaskrichel
_______________________________________________ RAS-run mailing list RAS-run@lists.openlib.org http://lists.openlib.org/cgi-bin/mailman/listinfo/ras-run
I'll be out of the bank all of next week. If you need me for anything, I suggest you wait a week. On Sun, Nov 13, 2011 at 7:39 AM, Ivan Kurmanov <duraley@gmail.com> wrote:
Christian and all, I've done some more work on ACIS & RAS. I've integrated most template (XSLT) changes that Christian did on RAS, both to screens and emails, into the main (Thomas') ACIS version. None of the RAS-specific changes went into the default template set (presentation/default/). I've created a new phrase.xml file based on the one I took from nebka.
Thomas may find these patches in the aras@holda.openlib.org:dev/acis.r repository, just as before.
Also, I've fixed the issues listed in the previous mail, see below. To me this version looks as if it is ready to be installed on RAS. Other problems may arise when we do this upgrade, but it is very hard to diagnose them now. So, I suggest to do the last round of testing and then install it. -ivan
On Sat, Nov 5, 2011 at 10:32 PM, Ivan Kurmanov <duraley@gmail.com> wrote:
Thanks to all who tested! The updated list of known issues: * http://ras.openlib.org/privacy-policy and some other pages do notwork, but http://ras.openlib.org/about and /faq work * profile-overview page returns 404 error * "click here" link on the affiliations screen * name variations on the research overview screen are shown with toomuch spacing between them
Anything else? -ivan On Fri, Nov 4, 2011 at 11:56 PM, Thomas Krichel <krichel@openlib.org> wrote:
In /affiliations
we have
Search institutions
click here
the anchor text is meaningless and document rather than subject talk.
"more about author affilitiations" would be better
Cheers,
Thomas Krichel http://openlib.org/home/krichel http://authorprofile.org/pkr1 skype: thomaskrichel
_______________________________________________ RAS-run mailing list RAS-run@lists.openlib.org http://lists.openlib.org/cgi-bin/mailman/listinfo/ras-run
_______________________________________________ RAS-run mailing list RAS-run@lists.openlib.org http://lists.openlib.org/cgi-bin/mailman/listinfo/ras-run
Christian, please do a test on http://ras.openlib.org/ and approve it, so that when Dan is back we could do the upgrade. -ivan On Sun, Nov 13, 2011 at 3:47 PM, Dan Hayes <dhayes501@gmail.com> wrote:
I'll be out of the bank all of next week. If you need me for anything, I suggest you wait a week.
On Sun, Nov 13, 2011 at 7:39 AM, Ivan Kurmanov <duraley@gmail.com> wrote:
Christian and all, I've done some more work on ACIS & RAS. I've integrated most template (XSLT) changes that Christian did on RAS, both to screens and emails, into the main (Thomas') ACIS version. None of the RAS-specific changes went into the default template set (presentation/default/). I've created a new phrase.xml file based on the one I took from nebka.
Thomas may find these patches in the aras@holda.openlib.org:dev/acis.r repository, just as before.
Also, I've fixed the issues listed in the previous mail, see below. To me this version looks as if it is ready to be installed on RAS. Other problems may arise when we do this upgrade, but it is very hard to diagnose them now. So, I suggest to do the last round of testing and then install it. -ivan
On Sat, Nov 5, 2011 at 10:32 PM, Ivan Kurmanov <duraley@gmail.com> wrote:
Thanks to all who tested! The updated list of known issues: * http://ras.openlib.org/privacy-policy and some other pages do notwork, but http://ras.openlib.org/about and /faq work * profile-overview page returns 404 error * "click here" link on the affiliations screen * name variations on the research overview screen are shown with toomuch spacing between them
Anything else? -ivan On Fri, Nov 4, 2011 at 11:56 PM, Thomas Krichel <krichel@openlib.org> wrote:
In /affiliations
we have
Search institutions
click here
the anchor text is meaningless and document rather than subject talk.
"more about author affilitiations" would be better
Cheers,
Thomas Krichel http://openlib.org/home/krichel http://authorprofile.org/pkr1 skype: thomaskrichel
_______________________________________________ RAS-run mailing list RAS-run@lists.openlib.org http://lists.openlib.org/cgi-bin/mailman/listinfo/ras-run
_______________________________________________ RAS-run mailing list RAS-run@lists.openlib.org http://lists.openlib.org/cgi-bin/mailman/listinfo/ras-run
Sorry for the delay. Very busy as always. Comments: research page: there is a [END-IF]--> in the right side bar I did a search for my documents. default should be on 'yes' for items corresponding to name variations, 'no' for the others. All are 'yes' currently. Or maybe put no choice by default? manual search, search by author: Died at /home/aras/acis/lib/ACIS/Web/Contributions.pm line 555. I think happens whenever there is a result. this offers a contact page, which is not valid: http://ras.openlib.org/contact_us Not sure my modifications were saved (I removed a slash in my homepage URL) The confirmation email for a new registration was not sent (but that may be a feature of the test site). On Tue, 15 Nov 2011, Ivan Kurmanov wrote:
Christian, please do a test on http://ras.openlib.org/ and approve it, so that when Dan is back we could do the upgrade.
-ivan
On Sun, Nov 13, 2011 at 3:47 PM, Dan Hayes <dhayes501@gmail.com> wrote:
I'll be out of the bank all of next week. If you need me for anything, I suggest you wait a week.
On Sun, Nov 13, 2011 at 7:39 AM, Ivan Kurmanov <duraley@gmail.com> wrote:
Christian and all, I've done some more work on ACIS & RAS. I've integrated most template (XSLT) changes that Christian did on RAS, both to screens and emails, into the main (Thomas') ACIS version. None of the RAS-specific changes went into the default template set (presentation/default/). I've created a new phrase.xml file based on the one I took from nebka.
Thomas may find these patches in the aras@holda.openlib.org:dev/acis.r repository, just as before.
Also, I've fixed the issues listed in the previous mail, see below. To me this version looks as if it is ready to be installed on RAS. Other problems may arise when we do this upgrade, but it is very hard to diagnose them now. So, I suggest to do the last round of testing and then install it. -ivan
On Sat, Nov 5, 2011 at 10:32 PM, Ivan Kurmanov <duraley@gmail.com> wrote:
Thanks to all who tested! The updated list of known issues: * http://ras.openlib.org/privacy-policy and some other pages do notwork, but http://ras.openlib.org/about and /faq work * profile-overview page returns 404 error * "click here" link on the affiliations screen * name variations on the research overview screen are shown with toomuch spacing between them
Anything else? -ivan On Fri, Nov 4, 2011 at 11:56 PM, Thomas Krichel <krichel@openlib.org> wrote:
In /affiliations
we have
Search institutions
click here
the anchor text is meaningless and document rather than subject talk.
"more about author affilitiations" would be better
Cheers,
Thomas Krichel http://openlib.org/home/krichel http://authorprofile.org/pkr1 skype: thomaskrichel
_______________________________________________ RAS-run mailing list RAS-run@lists.openlib.org http://lists.openlib.org/cgi-bin/mailman/listinfo/ras-run
_______________________________________________ RAS-run mailing list RAS-run@lists.openlib.org http://lists.openlib.org/cgi-bin/mailman/listinfo/ras-run
-- Christian Zimmermann FIGUGEGL! Economic Research Federal Reserve Bank of St. Louis P.O. Box 442 St. Louis MO 63166-0442 USA http://ideas.repec.org/zimm/
On Wed, Nov 16, 2011 at 6:53 PM, Christian Zimmermann <zimmermann@stlouisfed.org> wrote:
Sorry for the delay. Very busy as always.
Comments:
research page: there is a [END-IF]--> in the right side bar
fixed.
I did a search for my documents. default should be on 'yes' for items corresponding to name variations, 'no' for the others. All are 'yes' currently. Or maybe put no choice by default?
Thomas, any additional configuration needed for that?
manual search, search by author: Died at /home/aras/acis/lib/ACIS/Web/Contributions.pm line 555. I think happens whenever there is a result.
Sorry, this was me working on the thing, and actually leaving it in a half-broken state and going away. While I was doing my business, I realized you have a good chance of stumbling upon this. Anyway, fixed now.
this offers a contact page, which is not valid: http://ras.openlib.org/contact_us
Not sure my modifications were saved (I removed a slash in my homepage URL)
Not all of your modifications where saved, because some of them are hard or impossible to merge with new interface. Which homepage URL do you mean? Any other important modifications that are missing?
The confirmation email for a new registration was not sent (but that may be a feature of the test site).
I'll check that tomorrow. -Ivan
On Tue, 15 Nov 2011, Ivan Kurmanov wrote:
Christian, please do a test on http://ras.openlib.org/ and approve it, so that when Dan is back we could do the upgrade.
On Wed, 16 Nov 2011, Ivan Kurmanov wrote:
On Wed, Nov 16, 2011 at 6:53 PM, Christian Zimmermann <zimmermann@stlouisfed.org> wrote:
Sorry for the delay. Very busy as always.
Comments:
research page: there is a [END-IF]--> in the right side bar
fixed.
I did a search for my documents. default should be on 'yes' for items corresponding to name variations, 'no' for the others. All are 'yes' currently. Or maybe put no choice by default?
Thomas, any additional configuration needed for that?
manual search, search by author: Died at /home/aras/acis/lib/ACIS/Web/Contributions.pm line 555. I think happens whenever there is a result.
Sorry, this was me working on the thing, and actually leaving it in a half-broken state and going away. While I was doing my business, I realized you have a good chance of stumbling upon this. Anyway, fixed now.
I see, planting errors to see whether I would find them.
this offers a contact page, which is not valid: http://ras.openlib.org/contact_us
Not sure my modifications were saved (I removed a slash in my homepage URL)
Not all of your modifications where saved, because some of them are hard or impossible to merge with new interface. Which homepage URL do you mean? Any other important modifications that are missing?
I made changes to my profile, specifically to the URL of my homepage.
The confirmation email for a new registration was not sent (but that may be a feature of the test site).
I'll check that tomorrow.
-Ivan
On Tue, 15 Nov 2011, Ivan Kurmanov wrote:
Christian, please do a test on http://ras.openlib.org/ and approve it, so that when Dan is back we could do the upgrade.
-- Christian Zimmermann FIGUGEGL! Economic Research Federal Reserve Bank of St. Louis P.O. Box 442 St. Louis MO 63166-0442 USA http://ideas.repec.org/zimm/
Ivan Kurmanov writes
I did a search for my documents. default should be on 'yes' for items corresponding to name variations, 'no' for the others. All are 'yes' currently. Or maybe put no choice by default?
Thomas, any additional configuration needed for that?
yes. I tihnk the default is to use learning. I can try to look at this tomorrow. Cheers, Thomas Krichel http://openlib.org/home/krichel http://authorprofile.org/pkr1 skype: thomaskrichel
Thomas Krichel writes
Ivan Kurmanov writes
I did a search for my documents. default should be on 'yes' for items corresponding to name variations, 'no' for the others. All are 'yes' currently. Or maybe put no choice by default?
Thomas, any additional configuration needed for that?
yes. I tihnk the default is to use learning.
The way the software works now is taht it uses learning to order the above-me-propose-accept = 0.7 below-me-propose-refuse = 0.699999999 You can set these in the configuration. You can certainly bulid the "no choice" option by setting above-me-propose-accept = 1 below-me-propose-refuse = 0 But I believe this is suboptimal. A crude rule that comes from name variation matching is no longer supported, but it could be supported again with code changes. Cheers, Thomas Krichel http://openlib.org/home/krichel http://authorprofile.org/pkr1 skype: thomaskrichel
On Thu, 17 Nov 2011, Thomas Krichel wrote:
Thomas Krichel writes
Ivan Kurmanov writes
I did a search for my documents. default should be on 'yes' for items corresponding to name variations, 'no' for the others. All are 'yes' currently. Or maybe put no choice by default?
Thomas, any additional configuration needed for that?
yes. I tihnk the default is to use learning.
The way the software works now is taht it uses learning to order the
above-me-propose-accept = 0.7 below-me-propose-refuse = 0.699999999
You can set these in the configuration.
You can certainly bulid the "no choice" option by setting
above-me-propose-accept = 1 below-me-propose-refuse = 0
But I believe this is suboptimal.
A crude rule that comes from name variation matching is no longer supported, but it could be supported again with code changes.
Do I understand right that name variations are now useless?
Cheers,
Thomas Krichel http://openlib.org/home/krichel http://authorprofile.org/pkr1 skype: thomaskrichel
_______________________________________________ RAS-run mailing list RAS-run@lists.openlib.org http://lists.openlib.org/cgi-bin/mailman/listinfo/ras-run
-- Christian Zimmermann FIGUGEGL! Economic Research Federal Reserve Bank of St. Louis P.O. Box 442 St. Louis MO 63166-0442 USA http://ideas.repec.org/zimm/
'Christian Zimmermann' writes
Do I understand right that name variations are now useless?
No. The name variations is what is found. In the base installation thaht I have done, only those records that match name variations exactly are presented. What my base installation never had, (I don't know why) is a grouping of results by name variation types. (full names vs shortened names). This would be useless now since the ordering is done by learning. Note that libsvm and its utility scripts have to be installed. Cheers, Thomas Krichel http://openlib.org/home/krichel http://authorprofile.org/pkr1 skype: thomaskrichel
The learning is disabled now on the test install. Christian, do you want/need the learning? -ivan On Wed, Nov 16, 2011 at 11:37 PM, Thomas Krichel <krichel@openlib.org> wrote:
Ivan Kurmanov writes
I did a search for my documents. default should be on 'yes' for items corresponding to name variations, 'no' for the others. All are 'yes' currently. Or maybe put no choice by default?
Thomas, any additional configuration needed for that?
yes. I tihnk the default is to use learning.
I can try to look at this tomorrow.
Cheers,
Thomas Krichel http://openlib.org/home/krichel http://authorprofile.org/pkr1 skype: thomaskrichel
Tested again. Interesting that I was still logged in, and despite logging out, I am still able to get into the account without password. Not good. Strangely, I now have no suggestions. Doing a new registration, all are 'no', even exact name variations. Is that because learning is turned of? On Fri, 18 Nov 2011, Ivan Kurmanov wrote:
The learning is disabled now on the test install. Christian, do you want/need the learning?
-ivan
On Wed, Nov 16, 2011 at 11:37 PM, Thomas Krichel <krichel@openlib.org> wrote:
Ivan Kurmanov writes
I did a search for my documents. default should be on 'yes' for items corresponding to name variations, 'no' for the others. All are 'yes' currently. Or maybe put no choice by default?
Thomas, any additional configuration needed for that?
yes. I tihnk the default is to use learning.
I can try to look at this tomorrow.
Cheers,
Thomas Krichel http://openlib.org/home/krichel http://authorprofile.org/pkr1 skype: thomaskrichel
_______________________________________________ RAS-run mailing list RAS-run@lists.openlib.org http://lists.openlib.org/cgi-bin/mailman/listinfo/ras-run
-- Christian Zimmermann FIGUGEGL! Economic Research Federal Reserve Bank of St. Louis P.O. Box 442 St. Louis MO 63166-0442 USA http://ideas.repec.org/zimm/
'Christian Zimmermann' writes
Doing a new registration, all are 'no', even exact name variations. Is that because learning is turned of?
No, the default behaviour is that if there is no result to the learning, the suggestion is reject. Cheers, Thomas Krichel http://openlib.org/home/krichel http://authorprofile.org/pkr1 skype: thomaskrichel
'Christian Zimmermann' writes
Tested again. Interesting that I was still logged in, and despite logging out, I am still able to get into the account without password. Not good.
I think this is the default behaviour from the same IP. From a different IP? I don't think I changed in a module in a login part. Cheers, Thomas Krichel http://openlib.org/home/krichel http://authorprofile.org/pkr1 skype: thomaskrichel
participants (4)
-
'Christian Zimmermann' -
Dan Hayes -
Ivan Kurmanov -
Thomas Krichel