Semantic web - what can we do now
Literally, a wonderful article “ Elements of the semantic web ” has just skipped over the top. I started writing an answer to it, but it turned out to be too big, so I single it out into a separate topic.
So, most of the comments are divided into two categories: “Oh well, the ordinary user semantic markup of information is incomprehensible, inaccessible, and accordingly, he will not deal with it” and “Come on, everything will be fine, somehow we will get something of this bright future”.
I believe that those are right, and those, and we, those who consider the “semantic web” a desirable future, have opportunities to speed up its arrival.
')
PMSM, it is impossible to set as a goal the ubiquitous transition to the use of semantic technologies all at once, or in a short time. But it seems to me that it makes sense for enthusiasts to turn their attention to the developers and the tools they use. As soon as the possibility of (semi-) automatically adding semantic markup to the outputted information becomes cheap (in terms of labor costs, mental effort and time), many developers will begin to do this without even expecting immediate returns. That, in theory, should lead to the formation of a critical mass of semantic information on the Web and generate a boom in the writing of processors of such information, agents, etc. Look at the world of web services. Once Microsoft was pushing his SOAP hard, and now they won't talk to you anywhere, if you have never written SOAP or REST services. By the way, also an attempt to streamline information, and not entirely unsuccessful.
Many habraradi one way or another involved in the creation and publication of information on the web. The question of how this is a unique and meaningful information, while leaving aside. Let's say to the creation and publication of content. Question: how many people do it in the old manner, by hand? Those. open vim, emacs, or, God forbid, notepad.exe and stupidly write HTML pages, and then upload them via FTP to the sites? A bit, right?
For quite a long time, the vast majority of content before publication passes through one or another type of CMS - Content Management System. This also includes blog engines and other things. Or, if we ourselves write “site engines”, then we create such systems ourselves.
Most popular CMS are open source products. In principle, there are no obstacles to writing extensions for them that will make it easier for their users to create semantic markup for published information. Moreover, very many systems are already doing this anyway. The most primitive example is the <lj-user> tag in LJ. Many calendar systems use the hCalendar microformat (see the article " Elements of the Semantic Web ").
For the sake of interest, set yourself an extension to Operator Firewall and take a walk around the network - you will see a lot of interesting things.
Now imagine that in LJ, WordPress, Drupal, etc. There is a code that can respond to certain special tags (person, event, location), etc. Benefits for users - facilitating the formatting of their texts. The benefits to semantic network connectivity are enormous. I am now typing this text in a Habrareditor, and for links I use simple <a> tags. And if the "<a>" button in Habrareditor immediately inserted the microformat links (there is a special ontology on this topic, forgot, as they say, and too lazy to search now), then the semantic agent reading this page could link me to this article , the article to which I respond, references to literature, etc., thereby creating the basis for the machine analysis of this information.
Here you can swing even more, the possibilities are many. I will try to describe one example that I came up with in the process of writing this article.
I have written many web applications for my life on a bunch of Spring MVC technologies (or WebFlow) + Spring Framework + Hibernate + database. A fairly standard set in the world of Java. The general idea is standard - domain objects are in some way delivered to the presentation level, where they are transformed into HTML. The idea of a semantic extension for Spring was invented, which will consist of the following parts:
Please do not laugh out loud, I already wrote that the idea has literally just appeared, there is still to think and think.
So, I suggest that enthusiasts focus not on end users, but on development and publishing tools. The goal is to create some critical mass of semantic information on the Web. The benefits are enormous, and not only for monsters like Google. For us, the developers, this is an opportunity to be in the stream of a new revolution on the Web and all the connotations associated with this are self-realization, guaranteed jobs, a wide field for start-ups, etc. In my opinion, it's worth it.
So, most of the comments are divided into two categories: “Oh well, the ordinary user semantic markup of information is incomprehensible, inaccessible, and accordingly, he will not deal with it” and “Come on, everything will be fine, somehow we will get something of this bright future”.
I believe that those are right, and those, and we, those who consider the “semantic web” a desirable future, have opportunities to speed up its arrival.
Achievement of critical mass
')
PMSM, it is impossible to set as a goal the ubiquitous transition to the use of semantic technologies all at once, or in a short time. But it seems to me that it makes sense for enthusiasts to turn their attention to the developers and the tools they use. As soon as the possibility of (semi-) automatically adding semantic markup to the outputted information becomes cheap (in terms of labor costs, mental effort and time), many developers will begin to do this without even expecting immediate returns. That, in theory, should lead to the formation of a critical mass of semantic information on the Web and generate a boom in the writing of processors of such information, agents, etc. Look at the world of web services. Once Microsoft was pushing his SOAP hard, and now they won't talk to you anywhere, if you have never written SOAP or REST services. By the way, also an attempt to streamline information, and not entirely unsuccessful.
What are these tools?
Many habraradi one way or another involved in the creation and publication of information on the web. The question of how this is a unique and meaningful information, while leaving aside. Let's say to the creation and publication of content. Question: how many people do it in the old manner, by hand? Those. open vim, emacs, or, God forbid, notepad.exe and stupidly write HTML pages, and then upload them via FTP to the sites? A bit, right?
For quite a long time, the vast majority of content before publication passes through one or another type of CMS - Content Management System. This also includes blog engines and other things. Or, if we ourselves write “site engines”, then we create such systems ourselves.
CMS
Most popular CMS are open source products. In principle, there are no obstacles to writing extensions for them that will make it easier for their users to create semantic markup for published information. Moreover, very many systems are already doing this anyway. The most primitive example is the <lj-user> tag in LJ. Many calendar systems use the hCalendar microformat (see the article " Elements of the Semantic Web ").
For the sake of interest, set yourself an extension to Operator Firewall and take a walk around the network - you will see a lot of interesting things.
Now imagine that in LJ, WordPress, Drupal, etc. There is a code that can respond to certain special tags (person, event, location), etc. Benefits for users - facilitating the formatting of their texts. The benefits to semantic network connectivity are enormous. I am now typing this text in a Habrareditor, and for links I use simple <a> tags. And if the "<a>" button in Habrareditor immediately inserted the microformat links (there is a special ontology on this topic, forgot, as they say, and too lazy to search now), then the semantic agent reading this page could link me to this article , the article to which I respond, references to literature, etc., thereby creating the basis for the machine analysis of this information.
For developers
Here you can swing even more, the possibilities are many. I will try to describe one example that I came up with in the process of writing this article.
I have written many web applications for my life on a bunch of Spring MVC technologies (or WebFlow) + Spring Framework + Hibernate + database. A fairly standard set in the world of Java. The general idea is standard - domain objects are in some way delivered to the presentation level, where they are transformed into HTML. The idea of a semantic extension for Spring was invented, which will consist of the following parts:
- @Annotations, allowing to put domain classes and their fields in accordance with elements of a certain ontology
- The repository of ontologies used (not quite sure yet what it is for, but instinct tells you what to do)
- JSP tag library that would draw the necessary values in the desired (semantic) form
- Maybe some other extension of the standard Spring-A class of ModelAndView, which would automatically recognize entities in the model and something semantic with them would do
Please do not laugh out loud, I already wrote that the idea has literally just appeared, there is still to think and think.
Private total, or "why do I need it"
So, I suggest that enthusiasts focus not on end users, but on development and publishing tools. The goal is to create some critical mass of semantic information on the Web. The benefits are enormous, and not only for monsters like Google. For us, the developers, this is an opportunity to be in the stream of a new revolution on the Web and all the connotations associated with this are self-realization, guaranteed jobs, a wide field for start-ups, etc. In my opinion, it's worth it.
Source: https://habr.com/ru/post/79371/
All Articles