Foros

will php-openisis be developed in the future?

Autor	Mensaje
Escrito en: 18. 04. 2004 [20:26]
roger@infomed.sld.cu Roger Peña Escob Autor del tema registrado desde: 31.12.1969 Entradas: 0	hi after read some of the docs of malete isis server, i have a some questions about the future of this module 1- will ever have a write support? 2- will ever have any more developed? 3- will be move to some php class like the one you can find in the pear tree? please braulio, tell us some kind of action plan for the future. may be you ask "why a write suport if malete will give you that more easy and in a better way" well, in order to use malete i have to migrate my CDS/ISIS db to the new openisis format, that is something that made me feel unconfortable. i know that the CDS/ISIS file format have serius limitation and if you need to surpass those limitation you need to modify the "standard" but will the ISIS comunitie adopt this new standard? can you klaus give us some of your thougth about it? thank you very much for any reply roger ---------------------------------------------------------------------- Nodo central de la red Infomed (http://www.sld.cu) Usuario linux: 97152 (http://counter.li.org) Miembro del grupo de coordinacion de LinuxCuba (http://www.linux.cu) ---------------------------------------------------------------------- ------------------------------------------------- Este mensaje fue enviado usando el servicio de correo en web de Infomed http://webmail.sld.cu ------------------------------------------ Posted to Phorum via PhorumMail
	Perfil
Escrito en: 19. 04. 2004 [05:17]
paul@malete.org Klaus Ripke registrado desde: 31.12.1969 Entradas: 0	hola a todos On Monday 19 April 2004 03:26, Roger Pe=C3=B1a Escobio wrote: > 1- will ever have a write support? why not? Only problem is, this would require full exclusive locking of a DB during the request. > 2- will ever have any more developed? Braulio? > 3- will be move to some php class like the one you can find in the pear > tree? Note that the openisis/php/Isis classes support with next to no change both the old OpenIsis 0.9.1 server and the new Malete based server. Especially write support could also be based on the 0.9.1 (see the openisis/perl/Isis.pm, which does that). > may be you ask "why a write suport if malete will give you that more easy > and in a better way" well, in order to use malete i have to migrate my > CDS/ISIS db to the new openisis format, that is something that made me feel > unconfortable. > i know that the CDS/ISIS file format have serius limitation and if you need > to surpass those limitation you need to modify the "standard" but will the > ISIS comunitie adopt this new standard? > > can you klaus give us some of your thougth about it? Actually there is no need to "adopt this new standard" in the sense of dumping the existing formats and tools. Rather the opposite is true: The 0.9.0/1 knows how to write a .MST and .XRF in "DOS" format, but neither writes those in "UNIX" format nor writes the traditional index files. The reason is that those formats are not well suited for locking on parallel writes (it is well known that WinIsis fails to do this correctly, so sometimes, if rarely, data corruption occurs). The new software, on the other hand, is able to do full exports, masterfile and index, to both the "DOS" and "UNIX" (a.k.a. "Bireme") format. The latter requires a compile switch and may still have bugs (somebody can point me to a download of a medium sized UNIX format database for testing?), however, basically is implemented. So, to again make that point clear: =46rom a CDS/ISIS point of view, the new Malete database format can be regarded as just a temporary internal format used while processing, like e.g. WinIsis uses a couple of files in the work dir during full index update - nobody should bother. It is just a matter of convenience to obscure this fact from users, like automagically calling an import, if there's a new CDS database, and doing e.g. a daily export together with backups from an authoring system. =20 We learned from Adalberto in Havana that Bireme is not going to bother with any CDS/ISIS standards, since they are internally using much different formats. Likewise Davide left no doubt, if not in Havana then in Eschborn, that they will completely revise the file format: http://icg.openisis.net/ICG/IsisTodayAndTomorrow/Isis_today_and_tomorrow_issues4.txt (from http://icg.openisis.net/z/view/ICG/IsisTodayAndTomorrow). I am in contact with Davide, and maybe they will adopt parts of our work. However, which file format they will finally choose, we are going to support it just like we are supporting any other format. To summarize, our message is: "CDS/ISIS is dead - long live CDS/ISIS". Finally some notes on the future development: The Malete/php/Isis classes are going to stay as they are, i.e. probably extended, but without incompatible changes. They provide a rather ISIS-like view on a record as a flat list. We are right now developing a DOM-like approach called MOM (you guessed it: it's the Malete Object Model) to support deeply structured records (basically anything XML can do) in a uniform approach based on the good old v24^a style. This will be just an alternative interface to the very same database and server. The initial implementation will be in Tcl, however, just like DOM, the approach is general enough to also be supported in PHP. saludos Klaus ------------------------------------------ Posted to Phorum via PhorumMail
	Perfil
Escrito en: 19. 04. 2004 [12:25]
rucks@bl.fcen.uba.ar Nicolas Rucks registrado desde: 31.12.1969 Entradas: 0	Klaus, if you want, we can put our books ISIS (unix) database somewhere for you to dowload. I think you can call it "medium-sized": some 35.000 records in 15 Mb for the .mst file. Write me off-list. Nicolas Rucks Referencia, Biblioteca Central, FCEN, UBA. Argentina. ----- Mensaje original ----- De: Klaus Ripke <paul@malete.org> Para: Roger Peña Escobio <roger@infomed.sld.cu>; Php-openisis <php@openisis.org> Enviado: Lunes, 19 de Abril de 2004 07:17 Asunto: Re: will php-openisis be developed in the future? hola a todos On Monday 19 April 2004 03:26, Roger Peña Escobio wrote: > 1- will ever have a write support? why not? Only problem is, this would require full exclusive locking of a DB during the request. > 2- will ever have any more developed? Braulio? > 3- will be move to some php class like the one you can find in the pear > tree? Note that the openisis/php/Isis classes support with next to no change both the old OpenIsis 0.9.1 server and the new Malete based server. Especially write support could also be based on the 0.9.1 (see the openisis/perl/Isis.pm, which does that). > may be you ask "why a write suport if malete will give you that more easy > and in a better way" well, in order to use malete i have to migrate my > CDS/ISIS db to the new openisis format, that is something that made me feel > unconfortable. > i know that the CDS/ISIS file format have serius limitation and if you need > to surpass those limitation you need to modify the "standard" but will the > ISIS comunitie adopt this new standard? > > can you klaus give us some of your thougth about it? Actually there is no need to "adopt this new standard" in the sense of dumping the existing formats and tools. Rather the opposite is true: The 0.9.0/1 knows how to write a .MST and .XRF in "DOS" format, but neither writes those in "UNIX" format nor writes the traditional index files. The reason is that those formats are not well suited for locking on parallel writes (it is well known that WinIsis fails to do this correctly, so sometimes, if rarely, data corruption occurs). The new software, on the other hand, is able to do full exports, masterfile and index, to both the "DOS" and "UNIX" (a.k.a. "Bireme") format. The latter requires a compile switch and may still have bugs (somebody can point me to a download of a medium sized UNIX format database for testing?), however, basically is implemented. So, to again make that point clear: From a CDS/ISIS point of view, the new Malete database format can be regarded as just a temporary internal format used while processing, like e.g. WinIsis uses a couple of files in the work dir during full index update - nobody should bother. It is just a matter of convenience to obscure this fact from users, like automagically calling an import, if there's a new CDS database, and doing e.g. a daily export together with backups from an authoring system. We learned from Adalberto in Havana that Bireme is not going to bother with any CDS/ISIS standards, since they are internally using much different formats. Likewise Davide left no doubt, if not in Havana then in Eschborn, that they will completely revise the file format: http://icg.openisis.net/ICG/IsisTodayAndTomorrow/Isis_today_and_tomorrow_iss ues4.txt (from http://icg.openisis.net/z/view/ICG/IsisTodayAndTomorrow). I am in contact with Davide, and maybe they will adopt parts of our work. However, which file format they will finally choose, we are going to support it just like we are supporting any other format. To summarize, our message is: "CDS/ISIS is dead - long live CDS/ISIS". Finally some notes on the future development: The Malete/php/Isis classes are going to stay as they are, i.e. probably extended, but without incompatible changes. They provide a rather ISIS-like view on a record as a flat list. We are right now developing a DOM-like approach called MOM (you guessed it: it's the Malete Object Model) to support deeply structured records (basically anything XML can do) in a uniform approach based on the good old v24^a style. This will be just an alternative interface to the very same database and server. The initial implementation will be in Tcl, however, just like DOM, the approach is general enough to also be supported in PHP. saludos Klaus ------------------------------------------ Posted to Phorum via PhorumMail
	Perfil
Escrito en: 19. 04. 2004 [12:41]
braulio Braulio Solano Rojas registrado desde: 16.05.2008 Entradas: 2	Buenas tardes... I just want to say that I would like very much to do it. I have the time, but I do not have a regular income since 7 or 8 months. I use my time on looking for projects or clients. Things are not looking good since then and not even right now... On Monday, April 19, 2004 5:17 AM, Klaus Ripke wrote: > On Monday 19 April 2004 03:26, Roger Peña Escobio wrote: >> 1- will ever have a write support? > why not? Only problem is, this would require > full exclusive locking of a DB during the request. I was aware of this. I spent a lot of time thinking how to do it. I liked very much the idea of using System V semaphores to implement a locking system. However this would only work in UNIX. Maybe POSIX semaphores would be a better solution. I believe Windows NT 4,5,++ implements part of the POSIX standard. However POSIX is still very inmature in a lot of platforms. My last decision was to use the stuff inside the Zend Engine. There is a mutex type (for mutual exclusion) and it abstracts the platform. That is where I was. >> 2- will ever have any more developed? > Braulio? As I said, I would like very much to do it. I have a lot of ideas for it. I want also to put classes into the extension that can talk to Malete. That way there would be support for standalone and server. It depends very much on how will be my future situation. I don't know the answer to that question. God? ;-D >> 3- will be move to some php class like the one you can find in the pear >> tree? I do not understand very well that question. I was adding the stuff to PECL (http://pecl.php.net). But, as you know, rigth now the development I was doing is stalled. >> may be you ask "why a write suport if malete will give you that more easy >> and in a better way" well, in order to use malete i have to migrate my >> CDS/ISIS db to the new openisis format, that is something that made me feel >> unconfortable. >> i know that the CDS/ISIS file format have serius limitation and if you need >> to surpass those limitation you need to modify the "standard" but will the >> ISIS comunitie adopt this new standard? >> >> can you klaus give us some of your thougth about it? > Actually there is no need to "adopt this new standard" in the sense > of dumping the existing formats and tools. Just to add my 2 cents here, I believe that there are no CDS/ISIS standards. Maybe "de facto" standards. The question is about formats. About formats Roger I recommend you to read http://www.faqs.org/docs/artu/textualitychapter.html. And if you have time read all the book (I am doing it). Maybe after reading that chapter you will like to dump some formats... > To summarize, our message is: > "CDS/ISIS is dead - long live CDS/ISIS". > Finally some notes on the future development: > The Malete/php/Isis classes are going to stay as they are, > i.e. probably extended, but without incompatible changes. > They provide a rather ISIS-like view on a record as a flat list. > We are right now developing a DOM-like approach > called MOM (you guessed it: it's the Malete Object Model) > to support deeply structured records (basically anything > XML can do) in a uniform approach based on the good old > v24^a style. This will be just an alternative interface to the > very same database and server. The initial implementation > will be in Tcl, however, just like DOM, the approach is > general enough to also be supported in PHP. Sounds like good stuff... Â¡MuchÃsima suerte! Bye! B. ------------------------------------------ Posted to Phorum via PhorumMail
	Perfil
Escrito en: 19. 04. 2004 [15:29]
roger@infomed.sld.cu Roger Peña Escob Autor del tema registrado desde: 31.12.1969 Entradas: 0	Mensaje citado por Braulio José Solano Rojas <braulio@solsoft.co.cr>: > Buenas tardes... > > I just want to say that I would like very much to do it. I have the time, > but I do not have a regular income since 7 or 8 months. I use my time on > looking for projects or clients. Things are not looking good since then and > not even right now... sorry about it braulio , i'm trying to imaging how sad and unhappy you can be .. i remeber that Klaus sometime, at the begining of Clabel proyect, say that maybe someone will be willing to support some kind of develope i didn't ask more because is very dificult to implement something like this for us but maybe for you is more ease what do you think, we need some good programers skill in clabel , and clabel need very much the php-openisis write support... > > On Monday, April 19, 2004 5:17 AM, Klaus Ripke wrote: > >> 2- will ever have any more developed? > > Braulio? > > As I said, I would like very much to do it. I have a lot of ideas for it. > I want also to put classes into the extension that can talk to Malete. That > way there would be support for standalone and server. It depends very much > on how will be my future situation. I don't know the answer to that > question. > > God? > > ;-D > > > >> may be you ask "why a write suport if malete will give you that more easy > >> and in a better way" well, in order to use malete i have to migrate my > >> CDS/ISIS db to the new openisis format, that is something that made me > feel > >> unconfortable. > >> i know that the CDS/ISIS file format have serius limitation and if you > need > >> to surpass those limitation you need to modify the "standard" but will > the > >> ISIS comunitie adopt this new standard? > >> > >> can you klaus give us some of your thougth about it? > > Actually there is no need to "adopt this new standard" in the sense > > of dumping the existing formats and tools. > > Just to add my 2 cents here, I believe that there are no CDS/ISIS standards. > Maybe "de facto" standards. The question is about formats. About formats > Roger I recommend you to read > http://www.faqs.org/docs/artu/textualitychapter.html. And if you have time > read all the book (I am doing it). Maybe after reading that chapter you > will like to dump some formats... thanks for the link i'm looking into it right now thanks you very much for your work, think that without your php-openisis work a lot of peoples would never did good stuf, you give them the oportunity to be happy, to feel usefull and we want you to be happy because of that. This is food, and good food, for the spirit, for the soul. thanks roger ---------------------------------------------------------------------- Nodo central de la red Infomed (http://www.sld.cu) Usuario linux: 97152 (http://counter.li.org) Miembro del grupo de coordinacion de LinuxCuba (http://www.linux.cu) ---------------------------------------------------------------------- ------------------------------------------------- Este mensaje fue enviado usando el servicio de correo en web de Infomed http://webmail.sld.cu ------------------------------------------ Posted to Phorum via PhorumMail
	Perfil