Proposal: DB object syntax change
Dear working group I would like to propose syntax change of the role attribute of the ROLE object in the RIPE DB. Bellow you can find all details. I would like to hear your opinion about this idea and ask for the support to allow RIPE NCC to implement it. ------------- Short description: I would like to propose to change the syntax of the role attribute of the ROLE object from <person-name> to <organisation-name>, where both <person-name> and <organisation-name> are defined in RIPE Database Update Reference Manual. I would like to emphasize that I don't want to change the syntax of the nic-hdl attribute of the ROLE object. Nevertheless, my proposal influence the automatic mechanism of choosing the nic-hdl attribute value of the ROLE object, while it is created with AUTO directive. Rationale: Abuse-c implementation have influenced business procedures of LIRs and have led to increased importance of ROLE object. Since ROLE object is mandatory in the process of proper abuse-c registration, LIRs could be interested in automatic creation of it. Right now, due to its syntax, the role attribute of the ROLE object has number of limitations and cannot accept bunch of characters and or whole labels/words which are natural parts of company names. As an example, it is impossible to create ROLE objects with one of the following role attributes: - Company Father, Son & Daughter Abuse Team - 4U Network Abuse Team - 123 Ltd. Hostmaster - WOW! Network Dept. and so on. Moreover, the original intention was for the ROLE object to represent business roles and not personal data (see: paragraph 1.2.17 of the RIPE Database Update Reference Manual). Taking this into account the <organisation-name> would be the natural choice since it both perfectly solves the problem and is already well defined and implemented. Possible problems: As for now, the nic-hdl attribute of the ROLE object is auto-generated and based on the value of the role attribute. Since <person-name> do not allow all-numeric labels, this mechanism is quite simple. However, similar mechanism exists for the auto-generated organisation attribute of the ORGANISATION object, where <organisation-name> allows this kind of labels. While this is significant change to syntax, I would like to emphasize that <person-name> is a subset of <organisation-name>. This means that even if some people may rely on the current syntax, this change would most probably not influence their procedures. ----------- Best regards, Piotr -- gucio -> Piotr Strzyżewski E-mail: Piotr.Strzyzewski@polsl.pl
Piotr, On Fri, 31 Jan 2014 15:37:34 +0100 Piotr Strzyzewski <Piotr.Strzyzewski@polsl.pl> wrote:
I would like to propose syntax change of the role attribute of the ROLE object in the RIPE DB. Bellow you can find all details.
This makes sense to me. Cheers, -- Shane
Dear Piotr,
I would like to propose to change the syntax of the role attribute of the ROLE object from <person-name> to <organisation-name>
I think this is a very good idea.
I would like to emphasize that I don't want to change the syntax of the nic-hdl attribute of the ROLE object. Nevertheless, my proposal influence the automatic mechanism of choosing the nic-hdl attribute value of the ROLE object, while it is created with AUTO directive.
Good point. Another issue will be the query of role objects by name. The new syntax would allow you to put in the name characters that cannot be part of the query (see [*] below). Querying the role objects by nic-handle will still be possible, of course. The problem is solved in a somewhat similar way with organisation objects: one cannot query the organisation by name only by org-id. Anyway, I find this a minor issue, that should not prevent us from implementing your proposal. Best regards, Janos [*] quote from "RIPE Database Update Reference Manual, Appendices, A1. Object Attributes": <organisation-name> is a list of a most 12 words, each at most 64 characters in length. Words can contain alphanumeric characters, asterisk, plus and minus signs, forward slash and backslash, dash, quotes, at sign, commas, dots, underscores, ampersands, exclamation marks, colons, semicolons, brackets and square brackets. while, quote from "RIPE Database Query Reference Manual, Appendices, A1.1 Query Errors": %ERROR:108: bad character in input: An invalid character was passed in the query. The only allowed characters are letters, numbers and -_:+=.,@/?' so for example ampersands, exclamation marks and brackets are not allowed in the query.
On Wed, Feb 05, 2014 at 10:52:25AM +0100, Janos Zsako wrote: Dear Janos
I would like to propose to change the syntax of the role attribute of the ROLE object from <person-name> to <organisation-name>
I think this is a very good idea.
I would like to emphasize that I don't want to change the syntax of the nic-hdl attribute of the ROLE object. Nevertheless, my proposal influence the automatic mechanism of choosing the nic-hdl attribute value of the ROLE object, while it is created with AUTO directive.
Good point.
Another issue will be the query of role objects by name.
The new syntax would allow you to put in the name characters that cannot be part of the query (see [*] below).
Querying the role objects by nic-handle will still be possible, of course.
The problem is solved in a somewhat similar way with organisation objects: one cannot query the organisation by name only by org-id.
Anyway, I find this a minor issue, that should not prevent us from implementing your proposal.
Best regards, Janos
[*] quote from "RIPE Database Update Reference Manual, Appendices, A1. Object Attributes":
<organisation-name> is a list of a most 12 words, each at most 64 characters in length. Words can contain alphanumeric characters, asterisk, plus and minus signs, forward slash and backslash, dash, quotes, at sign, commas, dots, underscores, ampersands, exclamation marks, colons, semicolons, brackets and square brackets.
while, quote from "RIPE Database Query Reference Manual, Appendices, A1.1 Query Errors":
%ERROR:108: bad character in input: An invalid character was passed in the query. The only allowed characters are letters, numbers and -_:+=.,@/?'
so for example ampersands, exclamation marks and brackets are not allowed in the query.
This is very good point. However when I tried to reproduce it I have got another error: $ whois 'test !*&[] test' %ERROR:101: no entries found While using https://apps.db.ripe.net/search/query.html search form I have got 'Warning! No objects matched your search criteria.' So, I created the ORGANISATION object with the org-name filled with 'test !*&[] test' (without quotes) and I was able to query it by name as described above. I know that documentation is currently outdated. For example <person-name> is defined as not allowing '"Dr", "Prof", "Mv", "Ms", "Mr", no matter whether they end with dot (".") or not', which is not true anymore. Taking this into account I could imagine that this %ERROR:108: is also outdated or changed in some way. Could this be possibly confirmed by someone from NCC? Piotr -- gucio -> Piotr Strzyżewski E-mail: Piotr.Strzyzewski@polsl.pl
Dear Piotr, On 02/05/2014 11:22 AM, Piotr Strzyzewski wrote:
This is very good point. However when I tried to reproduce it I have got another error:
$ whois 'test !*&[] test'
%ERROR:101: no entries found
While using https://apps.db.ripe.net/search/query.html search form I have got 'Warning! No objects matched your search criteria.'
So, I created the ORGANISATION object with the org-name filled with 'test !*&[] test' (without quotes) and I was able to query it by name as described above.
Thank you for your valuable input. This restriction has been relaxed during the reimplementation project. Currently, the restrictions on queries are: - maximum 1024 bytes; - represents an UTF-8 encoded string.
I know that documentation is currently outdated. For example <person-name> is defined as not allowing '"Dr", "Prof", "Mv", "Ms", "Mr", no matter whether they end with dot (".") or not', which is not true anymore.
This issue has been fixed and will be part of the upcoming release 1.72 of the RIPE Database Software.
Taking this into account I could imagine that this %ERROR:108: is also outdated or changed in some way. Could this be possibly confirmed by someone from NCC?
We are very sorry that you encountered outdated information. We are striving to keep our reference manuals up to date, and will make sure this is fixed shortly. Kind regards, Agoston Horvath Senior Software Engineer RIPE NCC
On Wed, Feb 05, 2014 at 04:08:38PM +0100, Agoston Horvath wrote: Dear Agoston
On 02/05/2014 11:22 AM, Piotr Strzyzewski wrote:
This is very good point. However when I tried to reproduce it I have got another error:
$ whois 'test !*&[] test'
%ERROR:101: no entries found
While using https://apps.db.ripe.net/search/query.html search form I have got 'Warning! No objects matched your search criteria.'
So, I created the ORGANISATION object with the org-name filled with 'test !*&[] test' (without quotes) and I was able to query it by name as described above.
Thank you for your valuable input.
This restriction has been relaxed during the reimplementation project.
Currently, the restrictions on queries are: - maximum 1024 bytes; - represents an UTF-8 encoded string.
Good to know. Thanks.
I know that documentation is currently outdated. For example <person-name> is defined as not allowing '"Dr", "Prof", "Mv", "Ms", "Mr", no matter whether they end with dot (".") or not', which is not true anymore.
This issue has been fixed and will be part of the upcoming release 1.72 of the RIPE Database Software.
Taking this into account I could imagine that this %ERROR:108: is also outdated or changed in some way. Could this be possibly confirmed by someone from NCC?
We are very sorry that you encountered outdated information. We are striving to keep our reference manuals up to date, and will make sure this is fixed shortly.
Please excuse me for my scepticism about that. I have been told about the inconsistency with <person-name> by one of the NCC staff members on 20th of December 2013. I can imagine that someone forgot about that piece of documentation during the re-developed of the software (which was done a year ago as I was told). However, one month later there have been released the new version of RIPE DB software and it seems that no-one took care about corecting the documentation. And, sad to say, it seems from my perspective that it was necessary to write about that inconsistency loudly at this mailing list to force someone to take care about that. :( May I suggest the in-depth inspection of RIPE DB documentation? It seems that we (the users) cannot be sure how many inconsistecies like those two are still there. Best regards, Piotr -- gucio -> Piotr Strzyżewski E-mail: Piotr.Strzyzewski@polsl.pl
Dear Piotr, The RIPE Database release 1.71 that was deployed to production this week was prepared in the beginning of December. Any issues reported afterwards that were not critical for the user were scheduled to be fixed in version 1.72, which we were working on at that time. Just before the last RIPE Meeting (RIPE 67) we did a major review of the update and query manuals, which solved many of the issues following on from the software re-development project. As announced during the RIPE Database Working Group session at RIPE 67, we will continue to make further improvements to the documentation as it is quite difficult to maintain in its current state. The first action, which we’ve already taken, is to better incorporate the documentation updates in our release procedure. For any new release, the related documentation is reviewed and updated. This ensures that any new features and changes are properly documented. This does not mean that the current documentation is flawless. With changes that slipped through a long time ago it will take some time to find them all. Please bear with us on this and you can expect more improvements related to the documentation in the near future. The RIPE NCC welcomes any suggestions from the community on how we can do this better and we are grateful that you report any inconsistencies you find to us. Regards, Denis Walker Business Analyst RIPE NCC Database Team On 05/02/2014 17:18, Piotr Strzyzewski wrote:
On Wed, Feb 05, 2014 at 04:08:38PM +0100, Agoston Horvath wrote:
Dear Agoston
On 02/05/2014 11:22 AM, Piotr Strzyzewski wrote:
This is very good point. However when I tried to reproduce it I have got another error:
$ whois 'test !*&[] test'
%ERROR:101: no entries found
While using https://apps.db.ripe.net/search/query.html search form I have got 'Warning! No objects matched your search criteria.'
So, I created the ORGANISATION object with the org-name filled with 'test !*&[] test' (without quotes) and I was able to query it by name as described above.
Thank you for your valuable input.
This restriction has been relaxed during the reimplementation project.
Currently, the restrictions on queries are: - maximum 1024 bytes; - represents an UTF-8 encoded string.
Good to know. Thanks.
I know that documentation is currently outdated. For example <person-name> is defined as not allowing '"Dr", "Prof", "Mv", "Ms", "Mr", no matter whether they end with dot (".") or not', which is not true anymore.
This issue has been fixed and will be part of the upcoming release 1.72 of the RIPE Database Software.
Taking this into account I could imagine that this %ERROR:108: is also outdated or changed in some way. Could this be possibly confirmed by someone from NCC?
We are very sorry that you encountered outdated information. We are striving to keep our reference manuals up to date, and will make sure this is fixed shortly.
Please excuse me for my scepticism about that.
I have been told about the inconsistency with <person-name> by one of the NCC staff members on 20th of December 2013. I can imagine that someone forgot about that piece of documentation during the re-developed of the software (which was done a year ago as I was told). However, one month later there have been released the new version of RIPE DB software and it seems that no-one took care about corecting the documentation. And, sad to say, it seems from my perspective that it was necessary to write about that inconsistency loudly at this mailing list to force someone to take care about that. :(
May I suggest the in-depth inspection of RIPE DB documentation? It seems that we (the users) cannot be sure how many inconsistecies like those two are still there.
Best regards, Piotr
On Thu, Feb 06, 2014 at 03:16:26PM +0100, Denis Walker wrote: Dear Denis
The RIPE Database release 1.71 that was deployed to production this week was prepared in the beginning of December. Any issues reported afterwards that were not critical for the user were scheduled to be fixed in version 1.72, which we were working on at that time.
Right now, another 1.5 month after your mail and 3 months after my initial raport I have read about the new release of RIPE Database. As pointed above, the release 1.72 should contain any issues reported in the meantime that were not critical for the user and due to that fact were not included in the 1.71 release. It seems that the mistake with person-name is still not corrected. That's why I was then so sceptical about this kind of statements. :( I have to admit that this error is probably still not corrected due to the fact that the patch was only applied to the whois-client/src/main/java/net/ripe/db/whois/common/rpsl/AttributeSyntax.java file (https://github.com/RIPE-NCC/whois/commit/09fbbab3676a39a905b7972822596f1f75f...) and has not been applied to the whois-client/src/main/java/net/ripe/db/whois/common/rpsl/AttributeType.java file. If this is the case, this means that there is a flaw in the testing procedure and no-one have checked if the patch (hope that only this one) is correct. Moreover - the community was assured that the mistake with description of Error 108 will be fixed shortly. It seems that it hasn't been fixed yet. Best regards, Piotr -- gucio -> Piotr Strzyżewski E-mail: Piotr.Strzyzewski@polsl.pl
Dear Pior, It indeed seems like we overlooked the second instance of the help text (AttributeType.java). I’ve reopened the issue and will task the team to fix it. You can track the progress here: https://github.com/RIPE-NCC/whois/issues/121 Thanks for your thorough investigation! :-) Kind regards, Johan Åhlén Asst. Manager Database RIPE NCC On 25 Mar 2014, at 21:42, Piotr Strzyzewski <Piotr.Strzyzewski@polsl.pl> wrote:
On Thu, Feb 06, 2014 at 03:16:26PM +0100, Denis Walker wrote:
Dear Denis
The RIPE Database release 1.71 that was deployed to production this week was prepared in the beginning of December. Any issues reported afterwards that were not critical for the user were scheduled to be fixed in version 1.72, which we were working on at that time.
Right now, another 1.5 month after your mail and 3 months after my initial raport I have read about the new release of RIPE Database. As pointed above, the release 1.72 should contain any issues reported in the meantime that were not critical for the user and due to that fact were not included in the 1.71 release.
It seems that the mistake with person-name is still not corrected. That's why I was then so sceptical about this kind of statements. :(
I have to admit that this error is probably still not corrected due to the fact that the patch was only applied to the whois-client/src/main/java/net/ripe/db/whois/common/rpsl/AttributeSyntax.java file (https://github.com/RIPE-NCC/whois/commit/09fbbab3676a39a905b7972822596f1f75f...) and has not been applied to the whois-client/src/main/java/net/ripe/db/whois/common/rpsl/AttributeType.java file. If this is the case, this means that there is a flaw in the testing procedure and no-one have checked if the patch (hope that only this one) is correct.
Moreover - the community was assured that the mistake with description of Error 108 will be fixed shortly. It seems that it hasn't been fixed yet.
Best regards, Piotr
-- gucio -> Piotr Strzyżewski E-mail: Piotr.Strzyzewski@polsl.pl
On Fri, Jan 31, 2014 at 03:37:34PM +0100, Piotr Strzyzewski wrote: Dear WG
I would like to propose syntax change of the role attribute of the ROLE object in the RIPE DB. Bellow you can find all details.
[cut] Since there was no objections and there was some support for this idea, can we move forward and implement this change? Best regards, Piotr -- gucio -> Piotr Strzyżewski E-mail: Piotr.Strzyzewski@polsl.pl
On 24/04/14 13:46, Piotr Strzyzewski wrote:
On Fri, Jan 31, 2014 at 03:37:34PM +0100, Piotr Strzyzewski wrote:
Dear WG
I would like to propose syntax change of the role attribute of the ROLE object in the RIPE DB. Bellow you can find all details. [cut]
Since there was no objections and there was some support for this idea, can we move forward and implement this change?
I see no objections to this. I think there was clear consensus in the WG. NCC, can we go ahead with this please? Nigel
Dear colleagues, Thank you for the confirmation on this issue. The RIPE NCC will be deploying release 1.73 of the RIPE Database software to the Release Candidate Test Environment (RC) early next week. This syntax change for the ROLE object will be scheduled for inclusion in the next release 1.74. This will be released to RC in late May or early June. Regards, Denis Walker Business Analyst RIPE NCC Database Team On 25/04/2014 10:28, Nigel Titley wrote:
On 24/04/14 13:46, Piotr Strzyzewski wrote:
On Fri, Jan 31, 2014 at 03:37:34PM +0100, Piotr Strzyzewski wrote:
Dear WG
I would like to propose syntax change of the role attribute of the ROLE object in the RIPE DB. Bellow you can find all details. [cut]
Since there was no objections and there was some support for this idea, can we move forward and implement this change?
I see no objections to this. I think there was clear consensus in the WG.
NCC, can we go ahead with this please?
Nigel
participants (7)
-
Agoston Horvath -
Denis Walker -
Janos Zsako -
Johan Åhlén -
Nigel Titley -
Piotr Strzyzewski -
Shane Kerr