Proxy Server and its Applicability

proxy-server-and-its-applicability

Lately, there seems to be a slight confusion re. proxy server technology. As technologies get integrated /converged and new platforms come up (say for eg. an integrated solution comprising of Proxy server + TMS + CAT tool + voice transcription + vendor mgmt. + invoicing solution + payment solution), all bundled together giving an impression that proxy server technology platform does it all or comprises of all of it. Naturally for such new-age integrated solutions, since it is a mix of all different technologies or business functions, sometimes LSP (localization service providers) users or enterprise product company business managers don’t find it intuitive or find it difficult to train the staff who are used to some original standalone technology/tool thereby giving an overall impression that the proxy technology stack itself is inconvenient or cumbersome or heavy. Below are some clarifications from us regarding the proxy server and its applicability. For any different views or further information or clarification about general proxy technology or about our product, feel free to reach out to us at info@linguasol.net...

When to use Proxy?

You have a web application or website which is largely data driven
Basically this situation is actually same as regular software application, so you will need to go through i18n and L10N process if you follow conventional approach. The situation get further complicated if the application is custom developed as this would require complete code refactoring with all cost loaded on single deployment along with the risk of disruption which switching from existing to refactored code. Proxy eliminates need for i18n, and allows you to jump directly to the translation repository creation step.
Your portal consists of multiple website (vertical)
Conventional method will require you to first get all websites (maybe different CMSs) localization ready, and then localize each one independently as per their own data structures processes, which can get complex with increasing number of such websites and languages. Proxy provides single point localization and homogeneous localization methods.
Your website uses data from multiple sources: both internal and external
Since it may be logistically difficult (or sometime impossible) to export data from such sources (e.g. hotel description fetched from 3rd party website or ticker feed), use of conventional method may not work. Since Proxy does not differentiate between the data sources, it can deal with any data source.
You use CMS which cannot be made localization ready and not ready/possible to shift to CMS having localization support
Proxy is obvious choice in such cases.
You simply don’t have or don’t want to deploy internal manpower for content localization process, and your policy does not allow CMS admin to be given to external agency
Same again – Proxy facilitates completely external solution with almost no involvement of internal IT team.
You have a legacy enterprise product and don’t have enough clients to justify the costs for entire localization or do not have internal manpower for entire content localization process
Proxy facilitates completely external solution with almost no involvement of internal IT team or no modification to internal source code.

Proxy is not recommended for the following cases

Small, mostly static websites
Proxy is an overkill for such requirements. Small websites generally use Open Source CMS (WordPress, Joomla, Drupal), and either a simple mirroring or Import/Export Plugin should be good enough for the job on hand. If you are still looking for external solution as a shortcut (maybe since your vendor can’t support mirroring or import/export), and don’t bother too much about performance & SEO, then you can have a look at Scripting method (bablic.com, localizer.com)
Static contents, CMS with localization capabilities & import/export plug-in
If your website contains mostly static contents, and is implemented using a CMS which is localization ready (with all required capabilities) and reliable import/export plugin is available, then that is the way to go. Although you can still use Proxy in such situation, but it will not provide you with any major differentiator.
Global organization having independent regional websites
Since each reginal website contents are created and maintained independently, it is likely to be manually driven pick and choose translation of contents from main website. Proxy does not fit into such scenario.

Myths and facts

Myth: Proxy includes and replaces TMS
1. Proxy does not carry out or include content translation. Proxy technology does not include TMS or CAT tools. Proxy provides contents (and delta of the contents) to TMS for translations. The role of LSP continues to be same irrespective of use of Proxy or conventional method.
2. You can make use of any LSP/TMS/CAT and still follow the same process for ensuring translation quality and timelines. Proxy does not eliminate or replace this role. Ideally Proxy (technology) + LSP (TMS, CAT) make a team for providing localization solution.
3. Contents provided by Proxy are generally simpler to handle since these are flat files, and not data tables having complex relationship, which is normally the case with CMS export, and which also makes import high skill task. Hence, management of Proxy generated contents is easier, less error prone and has lesser person dependency.
4. The fear of “buggy memory export” could have been due to specific instances (maybe Proxy under development) – Proxy systems have become mature over time.
Myth: Proxy lack terminology management, translation memories, glossaries etc.
1. Same as above: Proxy technology does not include TMS, Terminology Management, Glossary Management or CAT tools. All these tasks are handled by LSP.
2. Some of the Proxy vendors also offer Language Services, and this might have led to confusion about the role of Proxy.
3. Most serious Proxy providers are technology driven companies and normally do not get into the area of language services. These two are actually complimentary roles.
Myth: Proxy is mostly used as content extractor
1. Proxy is not for content extraction. Unfortunately the misconception has got created that the primary task of content extraction is carried out by proxy.
2. Content extraction is carried out by Crawler. Normally included with Proxy infrastructure, the new age crawlers include schedulers for unassisted crawling, and can be configured to carry out frequent content change detection for specific pages and deep crawl normally scheduled for off-peak days/hours.
3. Crawler works in browser emulation mode, and hence is able to capture the text “as seen” by the real user. Crawling against test/staging server ensures that your translations are updated before contents are release on production server.
4. Proxy still possesses ability to collect untranslated text, but this is used more as exception – crawling is used for routine change detection. Crawling for proactive change detection and Proxy for reactive change detection ensure that no text remains untranslated and avoid any bleed-through.
5. Crawling suite now also include specific tools for collection of data from Interactive pages (which cannot be crawled automatically), and can be combined with automated crawling. Such tools enable quick implementation of Proxy with web application and portals having dynamic pages.
6. All these tools now capture not just the HTML tag data but also JSON/XHR and other resources like images, media, PDFs, JavaScript files to help “complete’ translation.
Myth: Bleed-through is inherent with Proxy
1. Combination of proactive change detection (using crawler with staging server) and reactive change detection (using Proxy for any residual change content) ensures that no changed content remains undetected & untranslated. See the above section for details.
2. Older method of relying only on Proxy did result in bleed-through, but crawler approach has addressed this problem.
Myth: You end up with incomplete translations if you use Proxy
1. New age Proxy suite can effectively locate, extract and replace text in Script tags, JS, JSON/XHR and even dynamic data having proprietary protocol.
2. All this is supported without any need to change the source code or manual addition of any markers. Specialized algorithms help locate the text inside Script tags and JavaScript files and put the markers. Proxy run-time uses these markers to locate the text in run-time data stream and carry out extraction-replacement.
3. The Proxy suite now have become so smart that it is capable of handling not just plain text inside the Script tags but also data in more complex formats like XML, JSON, CSV etc.
4. Moreover, the Translation Repositories now use strings with variables, and hence are capable of translating strings having combination of fixed and variable text – such instances are common in dynamic pages (e.g. discount percent in eCommerce sites).
5. This also largely addresses the issue of translation of Alerts & Error messages hidden inside Script tags in HTML. Generally the alerts are either hard-coded inside the Script tags or sourced from external list (which can be directly included in Translation Repository). Either way, Proxy can now translate Alerts/Messages.
Myth: Need help from customer’s Developer team to modify code or create markers
1. Earlier the Proxy solutions were able to process text in HTML tags but needed code modification to translate Script tag code as well as JavaScript files.
2. With the capabilities discussed in above section, the New age Proxy can handle all of these without need to make any change to the source code or add any markers.
3. Hence, no code change help is required from the developer team.
4. This reduces implementation time- you don’t need to schedule & deploy developers and there is hardly any testing/disruption as there is no code change required.
Myth: Translated Content are with Proxy provider only
1. No, the translated contents can be with the customer. Just the translation repository need to be published to the proxy server.
2. The Proxy itself can be deployed within customer’s data centre/hosting provider, so not just the contents but even the Proxy is within complete control of customer’s IT team. The translation are not only owned by the customer but also are in complete control & possession of customer.
Myth: Regional / Local contents cannot be created
1. Proxy solution had capability for URL replacement since long time.
2. New age Proxy provides facility for Conditional replacement of URLs.
3. With this, different page can be served depending upon the region or local requirements.
4. Actually the problem is exactly opposite – lots of Global brands are using regional instances of CMS and now face problem of dealing with lots of regional contents which are out of sync with the global website. This practice of independent creation of regional contents is inherently incapable of enforcing availability of updated content and end up sending different messages to the user. Many Global enterprise are trying to consolidate the contents under single authority to overcome the problem created due to too much of regional contents.
Myth: Proxy does not facilitate selective display of specific section
1. Proxy allows suppression of specific sections (blogs, reviews).
2. This can be implemented globally across the website using ClassIDs/ControlIDs, and can even be made URL specific.
3. In fact Proxy enables fetching translations from different sources depending upon Class/Tag type e.g. user review can be configured to be translated using MT.
Myth: Proxy does not support other Media (video), PDF etc. through links
1. Proxy facilitates automated URL provisioning & replacement depending upon language selection and path of translated media resources.
2. Translation of such media is always carried off-line. Proxy does not play role in extraction of text from such media. This is no different from the conventional method – no tools exist which allow reliable extraction of text from varied media or scanned PDFs. Even after extraction, the media may have to go through various human processing e.g. voice dubbing.
3. So, this is not a limitation of Proxy – it is at par with conventional method.
4. This feature has existed for long time now.
Myth: Proxy does not support SEO
1. Provisioning of regional/language sub-domains/sub-folders along with configuration of sitemap.xml brings Proxy URLs at par with static pages i.e. instruct search indexing bots to request pages for each language separately.
2. New age Proxy also includes two-way remapping of URLs to ensure that the search indexing bots see unique, keyword-rich, localized URLs.
3. Proxy crawler extracts Meta tags separately, and facilitates language specific fine-tuning of keywords.
4. All these put together bring Proxy at par with CMS solutions for SEO.
Myth: Proxy is for entire website only and cannot be implemented for specific section
1. Proxy provides you with list of URLs for translation. Pages that don’t need translation can be simply excluded.
2. You need not make this selection one-by-one for each URL. Proxy includes RegEx filter for URL inclusion / exclusion.
Myth: Estimation & hence budgeting takes too long
1. Proxy does not come into picture for estimation.
2. Estimation is carried out with the help of crawlers. Crawling includes detailed URL list and the page text (including JSON text).
3. Crawler can be configured to exclude specific types of URLS (specific strings or RegEx), and can be executed in multi-PC-multi-instance mode. Page crawling speeds of 20,000 pages/hour are easily available.
4. You can normally get word-count estimate inside a day even for very large websites (100,000 pages).
5. URL filtering can also be carried out post crawling, selection of URLs for translation as pre budget fitment becomes easy. In fact such exercise is way too complicated with conventional approach.
Myth: Proxy is expensive
1. It is a myth that Proxy pricing is always on page view basis. Different pricing models also exist e.g. Monthly/Quarterly subscription mode or even long term license.
2. The subscription / license model can be chosen as per the specific requirement.
3. Proxy can be deployed either on customer’s cloud or on external hosting. This provides flexibility in terms of hosting pricing.
4. Even in subscription mode, detailed analysis of language and URL access history is available, which allows you to get precise idea about the access frequency for sections / languages thereby facilitating further business decisions.
Myth: you get stuck with specific Proxy vendor
1. It is generally accepted that whenever you use a third party software, you are “stuck” with the software for some time (eg. MS Office for office tools or say Quickbooks for accounting and so on). Normally you purchase software after due-diligence, and actually don’t want to be “changing” the software very frequently. Same is true with Proxy or with CMS.
2. Your dependence on third party software cannot go away unless you develop every piece of software internally. But actually most customers don’t want to do that – what they want is to really outsource as much as possible (anything which is not a core business). Proxy really scores big on this as it is completely external, and hence assists customer in the process of smooth outsourcing.
3. The translation contents (which is really the IP) are anyway owned by the customer, so the customer retains the flexibility to move to another Proxy solution (there are many available now) or move to conventional method. All the investment in translation is preserved.
Myth: No Big Buyers
1. Many large enterprise have switched to Proxy. This includes eCommerce, retail, Internet Banking etc.
2. Each Proxy vendor focus on specific regions and verticals, but cover almost all verticals and regions between them.
Myth: Proxy cannot provide multilingual search
1. New age Proxy solutions include advanced options for handling word based and progressive search.
2. This involves reverse translation of inbound data and use of word translation dictionaries.
3. The user experience for translated languages is almost same as that of original language.
Myth: Proxy does not support format transformation, currency
1. New age proxy support format transformation for date/time, numbers etc.
2. Proxy can theoretically support currency handling, but it is best left to the Business Logic module since regional/currency pricing is not just currency conversion but is also dependent on business strategies.
Plus: Custom processing of page data based on language / region
1. Some of the Proxy solution have a major add-on: it facilitates custom Code/RegEx to be executed on server based on region/language for specific/all pages. Since the Code/RegEx is completely flexible (similar to writing Macro), the possibilities are unlimited.
Plus: Proxy can inject code
1. Proxy allows you to create JavaScript code and inject it in selected/all pages.
2. Such code could be useful for various tasks like user behaviour analysis, conditional display of specific information etc.
Plus: Assistance for Meaningful Translation
1. Proxy crawler & other capture tools are able to capture text broken across tags (due to attributes) as single segment. This produces better translation since the translator no longer has to manipulate translations across multiple segments.
2. The text is captured with attributes and href URLs. With this, CAT tools can facilitate integrated translation wherein the text and URLs are kept together.
3. The crawlers also capture the webpage and hence complete web-page reference becomes available to the CAT tools to help translator see the context and provide better translations. In case of CMS export, such context is generally not available.

About LinguaSol

LinguaSol is a LangTech firm with innovative cutting edge technology solutions for the localization industry. Its flagship product Linguify is a proxy server solution that helps in localizing the UI of any enterprise or web software application in any language without the need to modify the source code or DB. Linguify is used by major nationalized banks, Govt agencies in India for more than 8+ years. Enterprise firms like Fujitsu and others too continue to use it successfully. LinguaSol is a “technology-solution-only” firm with clear hands-off from the language services. Services are managed by its group company Fidel Softech with in-house translation services at LinguaSol thereby a pure products /technology company.

Please fill the form below to get the entire PDF:


We have a solution to translate your website in different languages, call LinguaSol +912022953848 or email info@linguasol.net

Website Translation Solutions: Please call LinguaSol +912022953848 or email info@linguasol.net